CSIC 2014( December )
50/- ` ISSN 0970-647X | Volume No. 38 | Issue No. 9 | December 2014 9 | December 38 | Issue No. No. | Volume 0970-647X ISSN Cover Story Boundaries of Algorithmic Computing 7
Cover Story Algorithmic Computing – A Perspective 10 Research Front Stable Marriage – Algorithm and Variants 23 Article Security Corner Defi ning and Describing Security Corner An Overview of Next Generation Multilayer Approach for Safe A Case Study of Vaayda Firewalls (NGFW) 35 Social Networking 30 Bazaar 37
www.csi-india.org www.csi-india.org CSI Communications | December 2014 | 1
CSI Communications Contents
Volume No. 38 • Issue No. 9 • December 2014
Editorial Board Cover Story Articles Boundaries of Algorithmic Computing Challenges in Using Aadhar as Unique Chief Editor 7 Prof. David M Keil 28 Identity Number for Delivery of Dr. R M Sonar e-Government Services Algorithmic Computing – A Perspective Rajesh Sharma Editors 10 N.S. Narayanaswamy Dr. Debasish Jana Defi ning and Describing Multilayer Dr. Achuthsankar Nair Technical Trends 30 Approach for Safe Social Networking Algorithms to Restructure the Websites Nandakumar Edamana Resident Editor 12 for Effi cient Browsing Mrs. Jayshree Dhere Harpreet Singh and Parminder Kaur Practitioner Workbench Programming.Tips() » Algorithm – AIM (Analysis in Minutes) 33 Fun with Bitwise Operators in C 14 B. Raj Kumar and M. Chandrakumar Peter Programming Amitava Nag Algorithmic Computing: A Detailed Published by 15 Oriented Thinking about Procedures Programming.Learn(“R”) » Executive Secretary Prof (Dr.) D G Jha and Ms. Kimaya Ambekar Shine with Shiny of R !!! Mr. Suchit Gogwekar Umesh P and Silpa Bhaskaran For Computer Society of India Recommendation Engines – A Generic 34 21 Architecture Design, Print and Ms. Seema Rawat, Mr. Praveen Kumar, Security Corner Dispatch by Prof. Sunil Kumar Khatri and Dr. Balvinder Shukla Information Security » CyberMedia Services Limited 35 An Overview of Next Generation Research Front Firewalls (NGFW) Stable Marriage – Algorithm and Variants Samriti Gupta, Balvir Kumar and P. K. Khosla Dr. Meghana Nasre 23 Case Studies in IT Governance, IT Risk and Information Security » Generating Random Numbers and 37 A Case Study of Vaayda Bazaar their Applications in Computing 26 Dr. Vishnu Kanhere Srabani Mukhopadhyaya
Please note: CSI Communications is published by Computer Society of India, a non-profi t organization. Views and opinions expressed in the CSI Communications are those of individual authors, contributors and advertisers and they may diff er from policies and offi cial statements of CSI. These should not be construed as legal or professional advice. The CSI, the publisher, the editors and the contributors are not responsible for any decisions taken by readers on the basis of PLUS these views and opinions. Although every care is being taken to ensure Brain Teaser genuineness of the writings in this publication, 40 CSI Communications does not attest to the Dr. Debasish Jana originality of the respective authors’ content. Happenings@ICT © 2012 CSI. All rights reserved. 41 Instructors are permitted to photocopy isolated H R Mohan articles for non-commercial classroom use without fee. For any other copying, reprint or CSI Reports 42 republication, permission must be obtained in writing from the Society. Copying for other than personal use or internal reference, or of CSI Elections 2015-2016/2017 43 articles or columns not owned by the Society without explicit permission of the Society or the CSI News 45 copyright owner is strictly prohibited.
Published by Suchit Gogwekar for Computer Society of India at Unit No. 3, 4th Floor, Samruddhi Venture Park, MIDC, Andheri (E), Mumbai-400 093. Tel. : 022-2926 1700 • Fax : 022-2830 2133 • Email : [email protected] Printed at GP Off set Pvt. Ltd., Mumbai 400 059.
CSI Communications | December 2014 | 3 Know Your CSI
Executive Committee (2013-14/15) »
President Vice-President Hon. Secretary Hon. Treasurer Mr. H R Mohan Prof. Bipin V Mehta Mr. Sanjay Mohapatra Mr. Ranga Rajagopal [email protected] [email protected] [email protected] [email protected] Immd. Past President Prof. S V Raghavan [email protected]
Nomination Committee (2014-2015) Prof. P. Kalyanaraman Mr. Sanjeev Kumar Mr. Subimal Kundu
Regional Vice-Presidents Region - I Region - II Region - III Region - IV Mr. R K Vyas Mr. Devaprasanna Sinha Prof. R P Soni Mr. Hari Shankar Mishra Delhi, Punjab, Haryana, Himachal Assam, Bihar, West Bengal, Gujarat, Madhya Pradesh, Jharkhand, Chattisgarh, Pradesh, Jammu & Kashmir, North Eastern States Rajasthan and other areas Orissa and other areas in Uttar Pradesh, Uttaranchal and and other areas in in Western India Central & South other areas in Northern India. East & North East India [email protected] Eastern India [email protected] [email protected] [email protected] Region - V Region - VI Region - VII Mr. Raju L kanchibhotla Dr. Shirish S Sane Mr. S P Soman Publication Committee (2014-15) Karnataka and Andhra Pradesh Maharashtra and Goa Tamil Nadu, Pondicherry, [email protected] [email protected] Andaman and Nicobar, Dr. S S Agrawal Chairman Kerala, Lakshadweep Prof. R K Shyamasundar Member [email protected] Prof. R M Sonar Member Dr. Debasish Jana Member Division Chairpersons Dr. Achuthsankar Nair Member Dr. Anirban Basu Member Division-I : Hardware (2013-15) Division-II : Software (2014-16) Division-III : Applications (2013-15) Prof. A K Saini Member Prof. M N Hoda Dr. R Nadarajan Dr. A K Nayak Prof. M N Hoda Member [email protected] [email protected] [email protected] Dr. R Nadarajan Member Division-IV : Communications Division-V : Education and Research Dr. A K Nayak Member (2014-16) (2013-15) Dr. Durgesh Kumar Mishra Member Dr. Durgesh Kumar Mishra Dr. Anirban Basu Mrs. Jayshree Dhere Member [email protected] [email protected] Important links on CSI website » About CSI http://www.csi-india.org/about-csi Membership Subscription Fees http://www.csi-india.org/fee-structure Structure and Orgnisation http://www.csi-india.org/web/guest/structureandorganisation Membership and Grades http://www.csi-india.org/web/guest/174 Executive Committee http://www.csi-india.org/executive-committee Institutional Membership http://www.csi-india.org/web/guest/institiutional- Nomination Committee http://www.csi-india.org/web/guest/nominations-committee membership Statutory Committees http://www.csi-india.org/web/guest/statutory-committees Become a member http://www.csi-india.org/web/guest/become-a-member Who's Who http://www.csi-india.org/web/guest/who-s-who Upgrading and Renewing Membership http://www.csi-india.org/web/guest/183 CSI Fellows http://www.csi-india.org/web/guest/csi-fellows Download Forms http://www.csi-india.org/web/guest/downloadforms National, Regional & State http://www.csi-india.org/web/guest/104 Membership Eligibility http://www.csi-india.org/web/guest/membership-eligibility Student Coordinators Code of Ethics http://www.csi-india.org/web/guest/code-of-ethics Collaborations http://www.csi-india.org/web/guest/collaborations From the President Desk http://www.csi-india.org/web/guest/president-s-desk Distinguished Speakers http://www.csi-india.org/distinguished-speakers CSI Communications (PDF Version) http://www.csi-india.org/web/guest/csi-communications Divisions http://www.csi-india.org/web/guest/divisions CSI Communications (HTML Version) http://www.csi-india.org/web/guest/csi-communications- Regions http://www.csi-india.org/web/guest/regions1 html-version Chapters http://www.csi-india.org/web/guest/chapters CSI Journal of Computing http://www.csi-india.org/web/guest/journal Policy Guidelines http://www.csi-india.org/web/guest/policy-guidelines CSI eNewsletter http://www.csi-india.org/web/guest/enewsletter Student Branches http://www.csi-india.org/web/guest/student-branches CSIC Chapters SBs News http://www.csi-india.org/csic-chapters-sbs-news Membership Services http://www.csi-india.org/web/guest/membership-service Education Directorate http://www.csi-india.org/web/education-directorate/home Upcoming Events http://www.csi-india.org/web/guest/upcoming-events National Students Coordinator http://www.csi-india.org/web/national-students- Publications http://www.csi-india.org/web/guest/publications coordinators/home Student's Corner http://www.csi-india.org/web/education-directorate/student-s-corner Awards and Honors http://www.csi-india.org/web/guest/251 CSI Awards http://www.csi-india.org/web/guest/csi-awards eGovernance Awards http://www.csi-india.org/web/guest/e-governanceawards CSI Certifi cation http://www.csi-india.org/web/guest/csi-certifi cation IT Excellence Awards http://www.csi-india.org/web/guest/csiitexcellenceawards Upcoming Webinars http://www.csi-india.org/web/guest/upcoming-webinars YITP Awards http://www.csi-india.org/web/guest/csiyitp-awards About Membership http://www.csi-india.org/web/guest/about-membership CSI Service Awards http://www.csi-india.org/web/guest/csi-service-awards Why Join CSI http://www.csi-india.org/why-join-csi Academic Excellence Awards http://www.csi-india.org/web/guest/academic-excellence- Membership Benefi ts http://www.csi-india.org/membership-benefi ts awards BABA Scheme http://www.csi-india.org/membership-schemes-baba-scheme Contact us http://www.csi-india.org/web/guest/contact-us Special Interest Groups http://www.csi-india.org/special-interest-groups Important Contact Details » For queries, correspondence regarding Membership, contact [email protected]
CSI Communications | December 2014 | 4 www.csi-india.org H R Mohan President’s Message From : President’s Desk:: [email protected] Subject : President's Message Date : 1st December, 2014 Dear Members I am happy to share the good news that Prof. Mahabala Distinguished Chair in Computational Brain Research has been established on 28th Nov 2014 by Mr. Kris Gopalakrishnan, co-Founder of Infosys at Indian Institute of Technology (IIT) Madras where Prof. Mahabala had served in the faculty of Computer Science and Engineering for over two decades. He has been adjudged by the students as “the Best Teacher” between 1981-82. Prof. H.N. Mahabala is one of our illustrious Past Presidents and the recipient of the CSI Life Time Achievement Award during CSI-2012 in Kolkata. We are proud of him. This is the most befi tting way of saying thank you to a person who has contributed so much to the fi eld of Computer Science and Technology. The fl agship event of the year - CSI-2014, the 49th Annual Convention at JNTU Hyderabad on the theme “Emerging ICT for Bridging Future” – is only two weeks away. We are thankful to Shri Ravi Shankar Prasad, Minister for Communications & IT and Law & Justice, Govt. of India and Shri K.T. Rama Rao, Minister for Panchayat Raj and IT, Govt. of Telengana for agreeing to be the Chief Guest and Guest of Honour at the inaugural session of the convention on 13th Dec 2014. In the Valedictory session on 14th Dec 2014, Shri Y.S. Chowdary, BC deserves congratulation, not only for hosting the KSSC-2014 but also for Minister of State, Science & Technology and earth Science, Govt. of India and a full-quorum presence. This included OBs, RSC, SSC etc. Dr. TC Manjunath, Shri Bandaru Dattatreya, Union Labour Minister, Govt. of India will be our guests Principal, Dr Mala Patil, HoD/CSE and the entire team at HKBKE deserve to deliver addresses. On the fi rst day of the convention on 12th Dec 2014, CSI appreciation for their dedicated eff orts in making the event a grand success. SIG-eGov will be hosting the Knowledge Sharing Sessions on eGovernance and I am happy to congratulate Mr. H.C. Sridhar, Manager of CSI BC also presenting the CSI Nihiliant eGovernance Awards during the convention. for having completed 25 years of service in CSI, similar to CSI ED staff Apart from the tutorials on state of the art technologies, special sessions on Mr. S. Natarajan, Manager(Finance) and M. Gnanasekaran, Manager ”IT enablement of Public & Govt. Sector” and “Inner Management for Human (Administration) who had similarly good innings at the CSI. Excellence” by Swami Bhoomananda Tirtha will be of interest to the delegates. Mr. Pramit Maakoday, CSI Vice President, International Region (2012- In addition to three keynote addresses on “Digital India – Opportunities and 14) along with Balasubbaraman Guruswamy, of ex CSI BC from the US, have Challenges Ahead” by Dr. Avinash Chander, Scientifi c Advisor to Defense on behalf of CSI prioritized the faciliation and introduction of immigration- Minister & DG, DRDO; “Digital Literacy for Smart India” by Mr. P. Sudhakar, related initiatives which ultimately resulted in executive actions announced CMD, ECIL; “Role of IT in Defence Systems” by Dr. V.G. Sekaran, DG- by the White House recently on immigration reforms for highly-skilled MSS, DRDO, the six main convention sessions namely “Cloud Computing workers. Our sincere thanks for their eff orts which will benefi t at least Technologies & Strategies for Business”, “Cyber Security & Information 25,000 Indians in America, particularly women who make up the majority of Warfare”, “Enterprise Networks & Intelligent Data Centers”, “Smart Cities those on H4 dependent visas. & IOT”, “Technology Imperatives for Education & Healthcare” and “Mobile It is concerning to read newspaper reports about stressed out IT Computing & Social Networking” will feature addresses by eminent speakers professionals who feel stigmatized by those in their professional circles from industry and academia. Contributed Paper Presentations will take place as they seek psychiatric help. Many turn into mental wrecks because they simultaneously. On the 1st day, Along with statutory CSI meetings, after the do not know they could get help or are reluctant to seek medical referral AGM, service and recognition awards to our members will be presented. For for fear of a stigma. The stress levels, meanwhile, are on the increase right the updates on the programme schedule, pl. visit http://csi-2014.org/ I am sure from the college days to the transition to the work place, resulting in a lot that CSI-2014 will be a memorable event for all of us and I look forward to of other health-related problems. Recently, I had the opportunity to visit Art seeing many of our members. Since this convention is held during our Golden of Living Centre in Bangalore and Gayatri Tirth Shantikunj in Haridwar both Jubilee period, we have extended considerable discounts in the registration fee run programmes on Yoga, personality refi nement (they prefer this term over to facilitate the attendance of a large number of our members. development) which ultimately aim to help professional achieve work-life In the CSI-2014 convention and Golden Jubilee celebrations, we balance. I have requested for their help in organizing programmes for our will be honouring four of our distinguished fellows of CSI with Life Time members and their response was very positive. We will further explore how Achievement Awards. We congratulate the following four LTA awardees: to go about introducing this initiative to help IT professionals cope with stress. Dr. C.R. Chakravarthy, who has served CSI for over three and a half decades During my recent visit for the International Conference on Advances in various capacities and of late who is synonym with the event “– IT for in Computing, Communications & Informatics, organized by the College Defence”; Professor D.K. Dutta Majumder, a founding member of CSI and of Engineering Roorkee, I had the opportunity of meeting the OBs and few a recipient of the Norbert Wiener Award; Brig. S.V.S. Choudhary, an Indian members of the CSI Dehradun and the re-activated CSI Haridwar along with Army Senior Offi cer who lead the CSI during 1992-94; and Prof. DVR Vithal, Dr. N.S. Chaudhari, Chairman of CSI Nagpur and Dr. Hoda, Chairman of CSI an academician with over four decades of service at Osmania University in Div I. I am extremely happy to note that all of them are very enthusiastic in computer science teaching and whose foresight in forming Student Branches furthering the vision of CSI. We are getting requests from a few NGOs for in CSI having a lasting impact in CSI. Furthermore, about 50 Fellows of CSI – a support in their Software Development projects. With the help of CSI ED, record number in any convention – will participate and deliberate in the Think we would like to involve our student members in such projects that will also Tank Meeting on the Past, Present and Future of CSI. expose students in real time projects. I had the opportunity of being at the 28th CSI Karnataka State Student I request members to encourage the school children and college Convention (KSSC) hosted by the HKBK College of Engineering, Bangalore students to participate in the Essay Contest on the theme “Harnessing the during 6-7, Nov 2014. CSI BC under the leadership of our Fellow Dr. R. Power of ICT for our New Initiatives” announced by CSI Chennai chapter and Srinivasan launched the state-level student convention way back in 1987 IEEE CS & PCS Madras. For details, please visit http://goo.gl/FziCmK and had set the trend, now followed by CSI ED in organizing various state- Again, due to space constraints, I am unable to script my views on a few level student conventions. The KSSC-2014 had attracted over 500 student items though I had proposed to do so in this month. As I conclude, I request member delegates from various student branches in Karnataka, participating all to encourage your colleagues and contacts to become Life Members of in ten events including technical quiz, programming, web development, CSI, taking advantage of the limited time discount off er in membership fee debate, mobile app development, code debugging and paper presentation. which will end by Dec 2014. Mr. Abdul Hameed, Administrator, HKBKCE in his address, elaborated on I look forward to seeing you at Hyderabad in Dec 2014. the need of innovation in the competitive environment. Mr. Rajesh Nambiar, With best regards Academic Alliance Manager and Global Social Media Manager addressed on employability issues and presented an enthusiastic talk on social media H.R. Mohan and analytics. I briefl y outlined the current trends in ICT and highlighted President entrepreneurial opportunities that exist today for the young talents. The CSI Computer Society of India
CSI Communications | December 2014 | 5 Rajendra M Sonar, Achuthsankar S Nair, Debasish Jana and Jayshree Dhere Editorial Editors
Dear Fellow CSI Members,
The theme for this issue is Algorithmic Computing. By algorithm, Kumar Khatri and Dr. Balvinder Shukla of Amity University, Noida we conceptualize a sequence of actions to compute a solution to on Recommendation Engines – A Generic Architecture that helps in a functional problem. Some problems are tractable that can be intelligent choice predictions for prospective buyer on the web. solved computationally in polynomial time while there remain Our Research Front section is enriched with two important several intractable problems that are believed not to have a contributions. First one is titled Stable Marriage – Algorithm and polynomial time algorithm. For intractable problems, sometimes, Variants by Dr. Meghana Nasre of Indian Institute of Technology we devise a solution for a simpler or restricted version of the Madras, elaborating techniques of fi nding a stable match between problem or try to fi nd an answer that is highly probable to be two sets of elements under preferences. Second one is by Dr. right, or approximately correct. The Turing machine remains as Srabani Mukhopadhyaya of Kolkata Campus of Birla Institute of the basis of all studies in computability and effi cient algorithms, Technology Mesra titled Generating Random Numbers and their thus serving as foundation of mathematical computation. Applications in Computing, which presents practical applications of While algorithmic computing relies on mechanical automatic randomization and probabilistic techniques. transformation of input to output, computing with intelligent machines may have to consider environmental interactions too, Our Article section is having two contributions on varied topics. and this could be imagined as interactive computing. The fi rst one is by Mr. Rajesh Sharma of Telecommunication Engineering Centre, New Delhi on Challenges in Using Aadhar as Our Cover Story section is enriched with two lucid and thought- Unique Identity Number for Delivery of e-Government Services and the provoking contributions. Prof. David M Keil of Framingham State second one is by Mr. Nandakumar Edamana, a software developer University, Massachusetts, USA has elaborated on Boundaries studying B.Sc. Computer Science in a college run by Institute of of Algorithmic Computing. Prof. Keil emphasizes on models of Human Resource Development, Kerala on Defi ning and Describing algorithmic computing through fi nite automata with equivalence Multilayer Approach for Safe Social Networking. among propositional logic, logic circuits, lookup tables and loopless transition systems and interactive computing through extensions In our regular Practitioner Workbench column’s Programming. of Turing machine, random-access machine and mu-recursive Tips() section, Mr. Amitava Nag of Academy of Technology, West function. Prof. N S Narayanaswamy of Indian Institute of Technology Bengal explains Fun with Bitwise Operators in C Programming. Umesh Madras has presented Algorithmic Computing – A Perspective. In his P and Silpa Bhaskaran continue their write-ups on app development article, Prof. Narayanaswamy talks about algorithmic computing as as Shine with Shiny of R under Programming.Learn(“R”). a phrase that stands for the design of a Halting Turing machine with emphasis on logical proof of algorithmic behavior deduced in a . . . . algorithmic computing as a phrase that stands for the design mathematical logic framework through axioms and inference rules. of a Halting Turing machine with emphasis on logical proof of In the process, he touches upon the upper and lower bounds of algorithmic behavior deduced in a mathematical logic framework effi cient algorithms with examples of some well-known algorithms through axioms and inference rules. and role of reductions among tractable and intractable problems from a mathematical standpoint. Under Information Security section in Security Corner we have an The Turing machine remains as the basis of all studies in interesting anecdote on An Overview of Next Generation Firewalls computability and effi cient algorithms, thus serving as foundation (NGFW) by Ms. Samriti Gupta, Mr. Balvir Kumar and Mr. P K of mathematical computation. While algorithmic computing Khosla of Terminal Ballistics Research Laboratory Chandigarh, relies on mechanical automatic transformation of input to output, DRDO. Dr. Vishnu Kanhere of CSI-SIG on Humane Computing computing with intelligent machines may have to consider searches answer to Algorithmic Computing – Problem or Solution environmental interactions too, and this could be imagined as through A Case Study of Vaayda Bazaar under Case Studies in interactive computing. IT Governance, IT Risk and Information Security subsection of Security Corner. Technical Trends section of this month has four articles. The fi rst Dr. Debasish Jana, Editor, CSI Communications presents one is by Dr. Harpreet Singh of DAV University Jalandhar and crossword for those who want to test their knowledge in Dr. Parminder Kaur of Guru Nanak Dev University Amritsar Algorithmic Computing under Brain Teaser column. Mr. H R on Algorithms to Restructure the Websites for Effi cient Browsing, Mohan, President, CSI and Former AVP (Systems), The Hindu, elaborating website graph structure improvement for better Chennai brings us the ICT News Briefs at a glance under various browsing experience. The second one is by B Raj Kumar and M sectors in November 2014 under regular column Happenings@ Chandrakumar Peter of Periyar Maniammai University, Vallam, ICT. We have other regular features like CSI Announcements, CSI Thanjavur on Algorithm – AIM (Analysis in Minutes), presenting Reports and Chapter and Student Branch News. a collection of time and space complexities of several common Please send your feedback, comments and views about CSI searching and sorting algorithms. The third one is by Prof. D G Jha Communications magazine to [email protected] and Ms. Kimaya Ambekar of K J Somaiya Institute of Management Studies and Research, Mumbai on Algorithmic Computing: A Detailed With warm regards, Oriented Thinking about Procedures to present fundamental building Rajendra M Sonar, Achuthsankar S Nair, blocks of programming to engage algorithmic techniques. The Debasish Jana and Jayshree Dhere fourth one is by Ms. Seema Rawat, Mr. Praveen Kumar, Prof. Sunil Editors
CSI Communications | December 2014 | 6 www.csi-india.org Cover Prof. David M Keil Story Assistant Professor, Computer Science Department , Framingham State University, Framingham, Massachusetts, USA
Boundaries of Algorithmic Computing
Computing began with devices to drive xi yi A looms, tabulate census data, and convert A X B data from punch cards into print-outs. memory The old computing paradigm of batch computing, typifi ed by COBOL programs E that produced print-outs from punch Fig. 3: Role of state in interacti ve computati on Fig. 5: Indirect interacti on cards, was algorithmic. An algorithm by defi nition transforms a fi nite input into a of pairs of inputs and outputs. fi nite output in fi nite time (Fig. 1). Furthermore, most Wegner, one of the authors of the original computer-science curriculum adopted by Finite input Finite output user-based computation Program today is not only the Association for Computer Machinery, interactive, but is actually Curriculum 68. Fig. 1: An algorithmic computati on multi-stream interactive In the field of software development, design notations have evolved with paradigms of computing and software. An algorithm always computes a The behavior of an interactive For decades, the flowchart and the function that maps from inputs to outputs. computing system is characterized module hierarchy were the most widely Output is strictly dependent on the most not by a function from possible inputs used notations. In the 1990s, with the recent input. to possible outputs, but rather as a rise of object-oriented technology, a new Today, most computing that involves stream of pairs of inputs and outputs. notation evolved that incorporated the users follows an interactive paradigm, in interactive paradigm. Thus the Unified which input and output follow each other Modeling Language (UML) incorporates repeatedly, often many times per second, - a second paradigm away from strictly use-case diagrams, interaction and in which most inputs depend on prior algorithmic computing. When we use our diagrams, and class diagrams, outputs (Fig. 2). mobile phones, when among many others. Object-oriented technology incorporates the interaction stream of inputs we log on to a social network, or when a paradigm with the notion of messages Agent Environment passed to objects. It is not an accident stream of outputs robotic device operates, multiple streams of that, when computational challenges Fig. 2: An interacti ve computati on inputs and outputs evolved from batch computing to are fl owing as part graphical user interfaces, a revolution of the computation. of object-oriented analysis and design At each step of an interactive In Fig. 4, we see how multiple computing occurred to accommodate the change. computation, an algorithm is executed agents may each interact concurrently with In the fi eld of theoretical computer that computes a function. Unlike the multiple others. science, change has not taken the same function computed by a batch program, course; it has not followed the pattern this function maps, not from external input of software-development technology. to output, but rather from a pair consisting Today, textbooks still claim that the Turing Agents/processes of external input and state of the interactive machine (the standard model of algorithmic computing agent, to a pair that consists Input/output stream computing) captures all computing, such of external output and state of the agent. as all the operation of a personal computer. Thus, when we have a conversation and Fig. 4: Multi -stream interacti on User interfaces, disk storage, cloud storage, you say, “How are you?” my response will and social networks are not captured by the not depend only on your question, but also In multi-stream interaction, much of Turing machine. on my state. When we click on a screen the interaction may be indirect. In Fig. 5, The computational project of building button using a touch pad, the result on our computing agent A interacts directly with cars that drive themselves, now well under screen is not computed as a function of agent X and not with agent B (we may call way, is solved by interaction and not by our click only, but also as a function of the E, consisting of X and B, the environment of algorithms alone. No eff ective algorithmic state of the device we are using. A). But A and B may interact indirectly. This procedure can take a car from a university Input to an interactive computation indirect interaction via the environment building to a cinema. A map can help, but is a stream (x, in Fig. 3), and output (y) is is the main reason for the rise of social the travel task can only be accomplished a stream. The behavior of an interactive networks and much other online software. by interaction with the environment along computing system is characterized not Thus, we have two paradigm shifts. the way. by a function from possible inputs to Both were identifi ed in the 1990s by Peter Ten years ago, Dina Goldin and others possible outputs, but rather as a stream proved that a model of interaction, the
CSI Communications | December 2014 | 7 Persistent Turing Machine, solves problems 1 is constructed from fi nite automata. A not solved by the iterated computations of stack machine is a fi nite automaton with a Turing machine. Other researchers, such 0 pushdown stack storage attached. as Farhad Arbab, Jan Van Leeuwen, Jiri The languages accepted by fi nite Wiedermann, Robin Milner, and researchers 1 automata are called regular and those in open computational systems and 0 accepted by stack machines are called environments for multi-agent systems, have context-free. All regular languages are Fig. 6: Finite automaton worked to model interactive computation. context-free, but not all context-free It seems probable that this research will languages are regular. All the languages be taken forward, perhaps by readers of with 1 and that, optionally, may end with a accepted by logic circuits are regular. this article. The body of theoretical work series of zeroes. developed by Wegner, Goldin, and others in Finite automata are in turn built particular invites elaboration and extension. from simpler devices, equivalent to the Finite automata are in turn built from The boundaries of algorithmic propositional logic. These may take the simpler devices .... may take the form of lookup tables, logic circuits, or form of lookup tables, logic circuits, The computational project of building loopless transition systems. Their sets of or loopless transition systems. cars that drive themselves, now well possible inputs are fi nite, whereas there is under way, is solved by interaction and no bound on the length of the input of a In turn, Turing machines are equivalent not by algorithms alone. fi nite automaton. A truth table is a lookup table that is equivalent to any logic circuit to two-stack stack machines, or, or loopless transition system. alternatively, to fi nite automata with tape computing, as treated by scholars, may The following three models of storage attached. The Turing machines and be extended in two directions, and computation are equivalent. The logic the random-access machines accept a set some have done so. The traditional circuit (Fig. 7a) computes a boolean of languages called recursive or recursively understanding of computing extends from (true/false or 0/1) function from three defi nable; and compute a set of functions fi nite automata, through stack machines, to boolean inputs. The truth table (Fig. 7b) on natural numbers called mu-recursive. Turing machines. It is known that Turing shows equivalent results and is used in Recursion and induction, well- machines are of equivalent computing computation by a lookup process. The known in mathematics, are very deeply power, or expressiveness, to random- connected with computation and its access machines. Turing machines operate 1 limits. By induction, we can prove that with symbols on tape, whereas RAMs 0 an algorithm computes the function operate via a simple assembler-like we say it computes. By use of recursive language that accesses memory locations Fig. 7: Automaton that accepts the language 10* function defi nitions, we may analyze the by symbolic names. time performance of algorithms. Thus the Finite automata (also called “fi nite mathematics of set theory, induction, and state machines”) are transition systems loopless transition system (Fig. 7c) accepts recursion theory is part of the professional that accept or reject their inputs, strings that take it, symbol by symbol, into toolkit of software developers. depending on the state in which an input a circled state following the labels on the By showing that a function is not leaves the automaton. An example is in transition arrows. Each also corresponds recursively defi nable, we may show that Fig. 6. It accepts strings such as “11000” to the sentence ((a ∨ b) ∧ ¬ (a ∧ b)) in it is not algorithmically computable, and rejects strings such as “001”. The propositional logic. hence it is a problem that is not worth languages accepted by fi nite automata are The languages (sets of (a, b) input the slightest investment in solving for all themselves considered an alternative way sequences) accepted by the above cases. A program-checking program that to model computation. described models are all fi nite. An could tell, from source code, whether a Each model of computation is instance of the logic circuit, truth table, or given program goes into an infi nite loop associated with a set of languages. For the loopless transition system accepts only a on some input, would have great value to a fi nite automaton in Fig. 6, the language fi nite number of strings. software development enterprise. But that specifi cation could be as follows: 1*0(0* Just as fi nite automata may be problem has unfortunately been proven to | 11*0)*. What precedes a star may be constructed from logic circuits or lookup be unsolvable computationally. You may repeated zero or more times, and the tables, the more powerful model of potentially save investors the waste of vertical bar (for OR) denotes options. computation known as the stack machine very large amounts of money by showing A simpler two-state FA is in Fig. 7. It accepts exactly the strings that start a a b a xor b 0 0 0 0 b 1 0 1 1 The traditional understanding of 0 1 1 0 1 0 computing extends from fi nite 1 01 10 automata, through stack machines, 1 1 0 to Turing machines. Fig. 7a: Logic circuit Fig. 7b: Truth table Fig. 7c: Loopless transiti on system
CSI Communications | December 2014 | 8 www.csi-india.org them that a program they would like to Persistent Turing Machine in 1999. develop cannot exist. A PTM is a three-tape TM extended Lookup tables / logic circuits are at by the fact that its third tape is a work the lower boundary of algorithmic ... the mathematics of set theory, tape whose contents persist between TM computation, and Turing machines / computations. (The standard TM model induction, and recursion theory is random-access machines are at the part of the professional toolkit of requires that the tape contain only the TM’s input at the start of a computation.) upper boundary. software developers. The work tape serves to store the state of the PTM and may be updated The boundaries of algorithmic The Turing machine reads one at the end of each TM computation, computation have been broken in symbol from its input/output tape, then known as a macrostep of a PTM. A PTM computing practice, twice over, by standard looks up which state to enter next, and computation step is the execution of desktop user interfaces and by multi- which symbol if any to write to the tape, one TM computation; one algorithm. stream interactive devices such as smart as well as whether to move its read/write Unlike executions of algorithms, PTM phones and robotic systems. Much of the head to the left or right on the tape or to computations may be infi nite. corresponding theoretical work has yet to stand still. In Fig. 8, q0 and q1 are states, The observable behavior of a TM be done, and the corresponding signifi cant and in state q0 the TM writes a ‘1’ if it reads is a function from inputs to outputs. The results are yet to be established as part of a ‘0’ but writes a ‘0’ if it reads a ‘1’. The TM observable behavior of a PTM is a set of the shared knowledge of computer science. negates one bit of its input and stops. streams of (input, output) pairs of strings. References q It is straightforward to defi ne [1] Farhad Arbab. Reo: A Channel-Based 0 0 1 q1 extensions to the random-access machine Coordination Model for Component model and the mu-recursive functions, Composition. CWI Report SEN-0203, similar to the extensions applied to the TM 2002. 1 0 to derive the PTM, and to prove that these [2] D Goldin, S A Smolka, P Attie, and extensions are equivalent to the PTM. E Sonderegger. Turing Machines, Fig. 8: A Turing machine This work has not yet been formalized Transition Systems, and Interaction. or published, however. Equally ripe for Information and Computation 194(2): A TM continues transitioning among research is the area of multi-stream 101-128, 2004. its states, reading and writing symbols interaction. Results that can be easily [3] Dina Goldin and David Keil. Interactive from the tape, until it enters a state that established include that asynchronous Models for Design of Software- is designated as terminal (accepting interaction is more expressive than Intensive Systems. In Proc. FInCo2005, Edinburgh, 2005. or rejecting). Thus this TM, like the FA synchronous interaction, and that [4] Dina Goldin, David Keil, and Peter and the logic circuit, produces a yes/no simulating a PTM computation on an Wegner. An interactive viewpoint on answer, but it also produces a string on asynchronous multi-stream interaction the role of UML. In Keng Siau and Terry its tape at the end of its computation. We machine is impossible in the general case. Halpin, Eds, Unifi ed Modeling Language: may think of a TM computation as a series Lookup tables / logic circuits are Systems analysis, design and development of accesses to two lookup tables, one at the lower boundary of algorithmic issues, Idea Group Publishing 2001, telling what to write and the other telling computation, and Turing machines / pages 249–263. what state to enter next. random-access machines are at the [5] D Keil and D Goldin. Modeling Indirect The work of Goldin and others upper boundary. We have referred to Interaction in Environments for Multi- on models of interaction extends the two limitations of models of algorithmic Agent Systems. Proc. E4MAS, 2005. Turing machine in a way similar to the computation: they cannot compute [6] David Keil and Dina Goldin. Adaptation straightforward, minimal extensions functions that are not recursively and Evolution in Dynamic Persistent applied in constructing the theoretical defi nable, and they cannot solve Environments. In Proc. FInCo2005. hierarchy of models of computation problems of an interactive type such [7] Jan van Leeuwen and Jiri Wiedermann. described above. We may call the kind as driving from one place to another. The Turing machine paradigm in of computing that occurs between a We have also indicated that an upper contemporary computing. In B Enquist user and a computer, or two computers boundary of sequential interaction is set and W. Schmidt, Eds., Mathematics Unlimited – and Beyond, Springer, 2001. that each wait their turns, sequential by synchronous interaction, in that no [8] Peter Wegner. Why interaction is more interaction. Its model is a sequential- model of sequential interaction can solve powerful than algorithms. CACM 40 interaction machine. Goldin formalized the problems of asynchronous sequential or (5), 1997. sequential-interaction machine as the multi-stream interaction. n
Prof. David M Keil has taught at Framingham State University, USA, since 1997. He has acted as Director of Assessment for computer science since 2009. Additional responsibilities include as Search Committee chair in Spring 2001 and member of search committee in 2004 and 2013-2014. He was Acting chair in Spring 2000-Fall 2000 and 2003-2004 and Faculty coordinator for computer-science lab during 1997-1998. He has presented at workshops on theory and practice of open computational systems, evolutionary computation, environments for multi-agent systems, foundations of interactive computing, and teaching and assessment in computer science. His research interests include interactive models of computation, evolutionary computation, artifi cial intelligence, database theory and Kolmogorov complexity. He can be reached at [email protected]. About the Author
CSI Communications | December 2014 | 9 Cover N.S. Narayanaswamy Story Associate Professor, Dept. of Computer Science and Engineering, Indian Institute of Technology Madras Algorithmic Computing – A Perspective What can be computed is formalized in Pythagoras himself is supposed have the framework of a subject called Formal ... a mathematical proof, or a logical proof Languages and Automata Theory[1], which learned of the possibility of creating ...... can defi nitely be verifi ed by a person is part of every undergraduate Computer formulae for diff erent mathematical objects with sight, but can also be verifi ed by a Science curriculum. All those decision from the Egyptians and Babylonians machine that can perform inferences in a problems that can be solved by a Turing Machine are those that can be computed. framework of logic an algorithm, for his/her study of Data Algorithms are special kinds of Turing Structures and Algorithms will know that Machines that halt on all inputs. The fi eld a C program or psuedocode can be used to logic. So, the correctness of an algorithm of Design and Analysis of Algorithms describe algorithms. Those of us who have is acceptable only in a framework of logic. has been a central topic in the fi eld of written assembly programs in our study of That is why as students of algorithms, Computing for nearly half a century now. Microprocessors will also recognize that we write logical proofs. A simple proof Prior to that, starting with Pythagoras, assembly code represents algorithms. illustrating the inferences made in a logical procedures to compute certain values, framework can be seen by analyzing the and procedures to compute one more popular sorting algorithm called Insertion entity with a given property were It is important to distinguish between a Sort[2]. The logical statement that plays a known for nearly two thousand years. logical proof and a scientifi c proof central role in the correctness analysis of In particular, given a Pythagorean triple insertion sort is – If insertion sort is run an (32 + 42 = 52), Pythagoras came up with a input consisting of n keys, then after the I-th Correctness, Halting, and Effi ciency method to compute another Pythagorean iteration where I lies in the range 1 and n, The correctness of an algorithm is triple, and this could be thought of as the fi rst I elements are in sorted order. This sacrosanct, it must be logically proved. among the earliest recorded algorithmic logical statement is then proved using the It is important to distinguish between a methods. Pythagoras himself is supposed principle of Mathematical Induction. Let logical proof and a scientifi c proof[3]- for have learned of the possibility of creating us illustrate the importance and ubiquity example, Software Developers usually formulae for diff erent mathematical of logical proofs in algorithms by another deploy testing as a way of believing that objects from the Egyptians and example that each of will be familiar with. an implementation is indeed correct. In a Babylonians[3]. In the world of Algorithms, Let us recall the Minimum Spanning Tree scientifi c proof, a statement is believed to one endeavours to optimize the resources Problem for which two popular algorithms be acceptable as true, if it can be verifi ed used during the execution of an Algorithm. are the Prim’s Algorithm and the Kruskal’s independent of the properties of the verifi er. The aim of this article to present to the algorithm[2]. In both the algorithms the Most of us who have studied science in reader the diff erent landmark algorithms, main logical statement that is proved is – school will recognize that repeatability of an the techniques used, and a few glimpses If the current data that has been computed experiment is a very important parameter into the bottomless pit of problems is guaranteed to be part of a fi nal optimum for the experiment to the recognized as a which have algorithms (Halting Turing solution, then among all the edges that can be valid one. A scientifi c proof is similar. On Machines), and we desire effi cient added to the current data without violating any the other hand, a logical proof is one where algorithms. of the constraints, the any one of the edges of the behaviour of the algorithm is inferred least cost can be used to extend the data, and in a framework of mathematical logic is also guaranteed to be part of a fi nal optimum Algorithms are special kinds of Turing (one that a fi nite number of axioms and solution. Using any of the acceptable Machines that halt on all inputs inference rules to derive new statements). edges of least cost is referred to as Greedy While many students of the subject will Choice, and the proof of correctness is by be familiar with this idea, it is useful to An Algorithm is a set of statements that the principle of Mathematical Induction, give a small example. Let us consider the precisely defi nes a sequence of operations and from the Theory of Matroids. There statement that the sum of three angles, on each input, and one that eventually are many more mathematical concepts in degree, of a triangle is 180. A scientifi c halts on each input. The word Algorithm is that play a very important role in the proof of this statement will consist of believed to stem from the name of a Latin correctness of Algorithms: The Ellipsoid many triangles, methods for drawing them, Translation of a book by Al Khwarizmi, who method and the Simplex Method for Linear using a compass and ruler, the values of was a Persian mathematician, astronomer, Programming[5], The Ford-Fulkerson and the 3 angles, and then the sum. On these and Geographer. A book by him titled On Edmonds-Karp Algorithms for computing examples, any person who can have sight the Calculation with Hindu Numerals when a maximum network fl ow[2][5], and the can measure and verify the truth of the translated into Latin became Algoritmi de Miller-Rabin test[2] and the Agrawal-Kayal- claim. This is what constitutes a scientifi c numero Indorum (meaning Al-Khwarizmi on Saxena algorithm[4] for primality testing. proof. On the other hand, a mathematical the Hindu art of Reckoning) (see Wikipedia The fi rst of the above examples, Linear proof, or a logical proof is based on the article on Algorithms). Algorithms are programming is an analysis of optimization Euclid’s axioms of geometry that can represented using many notations- an algorithms on convex polyhedra, the defi nitely be verifi ed by a person with sight, undergraduate student of Computer second one, about Network Flows is a but can also be verifi ed by a machine that Science will be familiar with the fact that problem that is extremely important can perform inferences in a framework of a halting Turing Machine represents in Transportation networks and other
CSI Communications | December 2014 | 10 www.csi-india.org mathematical proofs in algorithms, and for primality testing runs in time which is 6 the last one on Primality testing involves 0(log2 (n)) algebraic operations, and this There are many more problems for advanced Algebra, and also involves an All is exponentially smaller than square-root which we do not effi cient algorithms, and Indian Team. Termination of an Algorithm of n operations. Again, as in correctness, the only ineffi cient algorithms that we or Halting is traditionally the part of the effi ciency requires mathematical know are ones that involve exhaustive correctness proof. In our two examples handles from a variety of fi elds. Using enumeration of candidate solutions on the correctness of Insertion Sort and simpler combinatorial techniques, we Minimum Spanning Tree algorithms, note have known effi cient algorithms for have got the defi nition of the permanent that the logical statements involve in-built fi nding shortest paths in graphs, fi nding easily by changing the defi nition of the termination conditions. In the Sorting case, minimum spanning trees, for sorting a determinant, it is a well-known fact that the range of values of I, lying between 1 and given set of keys. Sorting a given set of the determinant of a matrix can be very n, guarantees termination. In the Minimum n keys is even more interesting, as we effi ciently calculated by the famous Spanning Tree example, from the subject know that a comparison based sorting Gaussian-Elimination algorithm using to Graph Theory, we can prove that in a algorithm cannot use lesser than nlog invariant properties of the determinant fi nite number of iteration the set of edges (n) comparison operations. This is a that, again, we studied in school. However, that can be added without violating any lower limit on the effi ciency, an inherent the permanent function so easily obtained of the constraints will become empty, at problem specifi c limit. While design of by modifying the determinant function, which point, the algorithm stops. This is an effi cient algorithms have been enticing remains an elusive problem for algorithm appropriate place to stop this discussion on and exciting, the goal of proving lower designers, and is believed to be inherently correctness and halting, and this is in the bounds on effi ciency have been daunting. hard for effi cient algorithms. There are interests of using the space limit for this Our mathematical understanding of many more problems for which we do article effi ciently. algorithms are just evolving. Are there not effi cient algorithms, and the only The most ineffi cient algorithm is to problems for which we cannot do any ineffi cient algorithms that we know are enumerate each candidate solution or better than just do an exhaustive, ones that involve exhaustive enumeration almost all candidate solutions and select therefore ineffi cient, enumeration to of candidate solutions. Let us list a few- a solution when we fi nd one. While this solve a problem? Many. Let us start Boolean Satisfi ability, Travelling Salesman with the simplest, and one that a fresh Problem, and the Graph Colouring The most ineffi cient algorithm is to reader is likely to be unaware of- A Problem[6]. They and many other problems enumerate each candidate solution or society of n men and women have given that they can be reduced to (see[2] for more almost all candidate solutions and select a symmetric marriage preferences for about reductions) seem to be waiting for solution when we fi nd one each other. A symmetric preference is the Indian Ingenuity that introduced zero one in which a pair have agreed to be or shoonya to Mathematics. You are all may seem an exaggeration, there are married, if necessary. How many pairings welcome, there is so much to be explored. many problems that still only admit are there in the society? There is no such correct algorithms. Indeed, before effi cient algorithm known to solve this References [1] Hopcroft, John E; Motwani, Rajeev; the Agrawal-Kayal-Saxena algorithm problem, yet. Further, it is believed that Ullman, Jeff rey D. (2013). Introduction for primality testing in 2002, such an this problem cannot be solved in time to Automata Theory, Languages, and algorithm was the effi cient algorithm which is a polynomial in n. This problem Computation (3rd ed.). Pearson. that most of us knew. Let us visit this as is identical to computing the Permanent [2] Thomas H Cormen, Charles E Leiserson, a complete example (for many of us this of a square matrix, and research on this Ronald L Rivest, and Cliff ord Stein. 2009. was our fi rst programming exercise): take problem is being spearheaded by Indian Introduction to Algorithms, Third Edition a positive integer n as input (remember, n research groups headed by Agrawal of (3rd ed.). The MIT Press. is represented using log2 (n) bits), and test the primality testing fame. Permanent [3] Simon Sin gh. 1997. Fermat’s Enigma: The if it is a prime number. The algorithm that is a fasicnating function that looks very Epic Quest to Solve the World’s Greatest was immediately implemented was to test similar to its very famous cousin, so to Mathematical Problem. Walker, New York if any integer between 2 to square-root of speak, the determinant of a square matrix. [4] Agrawal, Manindra; Kayal, Neeraj; Saxena, Nitin (2004). “PRIMES is in P”. Annals of n was a factor. If a factor was found, then To get the defi nition of the permanent Mathematics 160 (2): 781–793. n was reported as composite, else it was from the defi nition of the determinant [5] Schrijver, Alexander (1987), Polyhedral reported as prime. How effi cient is this that we have all studied in our school Combinatorics, Centrum voor Wiskunde algorithm? The input is the number n, and days, replace all the negative signs in the en Informatica. we only perform square-root of n division determinant formula with a positive sign. [6] Garey, M R; Johnson, D S (1979). Victor operations. It looks effi cient. Actually, it is Remember that the signs of the matrix Klee, ed. Computers and Intractability: A Guide an exponential-time algorithm in the input entries are not the ones that you are to the Theory of NP-Completeness. A Series of size, which is log2 (n) bits to represent n. asked to change, it is the negative signs Books in the Mathematical Sciences. San So the Agrawal-Kayal-Saxena algorithm in the determinant formula. Now that you Francisco, Calif.: W H Freeman and Co. n
N.S.Narayanaswamy teaches at IIT Madras, after a B.Sc in Mathematics from the University of Madras, M.E in Electrical Engineering from IISc, and Ph.D from IISc. He likes to teach, and spends most of his time teaching courses involving Mathematics and Programming. He can be reached at [email protected] About the Author
CSI Communications | December 2014 | 11 Technical Harpreet Singh* and Parminder Kaur** *Assistant Professor, Department of Computer Science & Engg., DAV University Jalandhar Trends **Assistant Professor, Department of Computer Science & Engg., Guru Nanak Dev University Amritsar, India
Algorithms to Restructure the Websites for Effi cient Browsing Introduction new products and schemes is added • Outdegree limit- There should With the expansion of internet, the size regularly and information about old be a constraint on the number of of the websites is growing at a faster products is removed. hyperlinks out of a webpage. rate and the number of Internet users The methods to reorganise the • Link constraint- There are some links is also increasing exponentially. Many graph structure are further classifi ed into which should not be removed to fi rms and government agencies are following three approaches: maintain the service logics specifi ed providing services through applications • Mathematical Programming by website design. such as e-commerce and e-government Techniques- In these approaches, respectively. This has created a problem Algorithms for Website Restructuring the problem is formed as 0-1 This section briefl y discusses the various for the website designers to develop programming problem with some such a structure of the website so that categories of methods for website graph constraints to be satisfi ed. These structure improvement. Figure 2 shows users do not get disoriented and get approaches use tools (Lin, 2006) to satisfi ed with the website contents. the basic web graph reorganization model. solve the models. The navigation behaviour of users and Effi cient browsing refers to navigating to • Heuristic Methods- These methods are the relevant web page quickly with less the current web graph structure are fed based on heuristics such as combining to the model as input and the algorithm eff ort. It is a normal tendency of the web two nodes, removal or addition of users that they tend to leave a website if generates the improved graph structure links etc. These approaches do not as output. it takes long to reach to the target page. It produce the optimal solution and also has also been observed that the browsing take less time. behaviours of users change frequently. • Meta-Heuristic Techniques- Mostly, Hence algorithms to rearrange the metaheuristics are derived from structure of websites are needed. some natural process such as food Web Graph Structure foraging process of ants. These The structure of a website can be approaches use predefi ned strategy considered as a directed graph where to fi nd the near optimal solutions. Fig. 2: Website graph restructuring model nodes represent webpages and edges The websites which reorganize represent the hyperlinks between their graph structure according to the webpages. Figure 1 represents such a behaviour of the users are known as Heuristic Methods structure. adaptive websites. This restructuring can A very successful heuristic method be performed in two ways: Customization is developed by Fu et al. (2002). The and Transformation. In Customization, web webpages are classifi ed into two pages are generated which include links to categories: index pages and content pages another pages according to behaviour of a based on the page access information. particular user. Transformation approaches The general scheme of reorganization is include reorganizing the graph structure to reduce the navigation time by reducing of a website to improve the browsing the number of intermediary index experience for all the users. Both these pages a user has to browse to reach the approaches use web log data at the web destination webpage. To accomplish this, Fig. 1: Hyperlink graph example servers to get the browsing behaviour of frequently accessed pages are positioned users. Here, browsing behaviour of users higher up in the link structure, i.e. nearer refers to the webpage navigation patterns to the home page and pages that are not The websites are divided into two of web users. The website graph structure accessed frequently are placed lower categories: Informational or Static websites improvement models follow certain in the website link structure. Starting and Dynamic websites. Informational constraints. Some of the constraints are: from the home page, the webpages are websites refers to the websites whose • Connectivity constraint- This examined sequentially. For every page, contents remain unchanged for a long constraint defi nes that every other the predecessor and successor nodes are period of time. Examples consist the webpage should be reachable from considered, where a predecessor is any websites of hotels, universities and the home page. page that has a link to it and successor hospitals. Dynamic website refers to the • Depth constraint- It specifi es that nodes are pages that are pointed by the websites whose contents change regularly. there should be limit on maximum current page. There are diff erent cases E-commerce websites are examples of number of links to be followed to depending on the number of predecessor these websites where information about reach a particular page. and successor nodes and for each case,
CSI Communications | December 2014 | 12 www.csi-india.org diff erent actions may be taken according metaheuristic algorithms. Lin and Tseng solution. In this method, the original to the frequency and category of the (2010) have developed a model based on structure is used as the initial solution. The pages involved. ant colony system (Dorigo, Maniezzo, and structure is altered by two operations of Mathematical Programming Methods Colorni, 1996) for website reorganization. link insertion and deletion. A combination Lin (2006) has successfully applied the The model works in two stages. In the fi rst of a link deletion and a link insertion leads 0-1 programming model to optimize stage, the ant colony method is applied to a neighboring solution. The algorithm the web graph structure. Every link in to fi nd a spanning tree which follows the begins from an initial solution S which can the website link structure has an access depth and outdegree constraints. Here be generated by some heuristic method. ants move along the edges and keep on A set N(S) of the neighboring solutions of frequency fij obtained by using the web usage mining process (Lin, 2006) on depositing the pheromone on the edges. the current solution is considered and the the web logs stored at web servers. The One by one, edges are added to the move that improves most the objective spanning tree. When the construction function value f is selected. If there are no access frequency fij corresponds to the number of users moving from webpage i of the spanning tree is completed then improving moves, tabu search chooses to webpage j. The goal of optimizing the second stage starts. In the second stage, the move that least degrades the goal web graph structure is to maximize the the edges with largest weights are function. This procedure is repeated until frequency summation of all the links with continuously added to complete the graph a stopping condition is reached. certain constraints.The objective function structure that satisfi es the depth and Wang, Wang and Ip (2006) worked of model is described below. outdegree constraint. in the area of dynamic websites and Saremi, Abedin, and Kermani (2008) developed a method to optimize the link ∑ have developed a method in which the structure for E-supermarket websites. fij xij Website Structure Optimization problem Hopfi eld networks are a special case of Max (f,f)∈E (1) is modeled as quadratic assignment (QAP) Artifi cial Neural Networks and have been (Loiola et al., 2007) type problem. Ant a successful method for combinatorial The parameter xij denotes the presence of colony meta-heuristic technique (Dorigo, optimization problems. The basic link (i, j). If xij = 1, it means that a hyperlink Maniezzo, and Colorni, 1996) is employed elements of neural network are neurons exists from page i to j. If xij = 0, it denotes to solve the problem. The QAP is formulated and their connections. Hopfi eld networks that page j cannot be accessed directly as facility location problem (Loiola et al., are constructed from artifi cial neurons. from i. 2007). The facilities location problem is These artifi cial neurons have N inputs If the web graph structure of a about fi nding the better or best location and a weight w is associated with every website is changed signifi cantly then the of facilities when the distances between input. Four kind of neural units have been old users may fi nd it diffi cult to browse locations and the demand fl ows among used in the technique. One kind of neural the website. To address this issue, the facilities is given. It can also be defi ned unit represents the presence or absence Chen and Ryu (2013) have developed a as the problem of fi nding a minimum cost of a link in the web structure. Other three mathematical programming model that allocation of facilities into locations. The kinds of neural units represent the three assists the user navigation on a website. web pages are considered as facilities and constraints related to indegree, outdegree Here the changes made to the web graph the connectivity (Saremi, Abedin, and and total number of links respectively. structure are minimized. In this model, the Kermani, 2008) between pages is taken as Authors observed that Hopfi eld network out-degree is taken as a cost term in the the fl ow between the facilities. The shortest technique is suitable for web graph of objective function and webpages that have distance between two pages is taken as the less than one hundred nodes. Recently an more links than the specifi ed threshold are distance between locations. The problem extension of the approach developed by penalized, hence the out-degree of a page is to allocate diff erent web pages to the Lin and Tseng (2010) has been developed may be more than the threshold if the cost possible locations in a website based on by Singh and Kaur (2014).The method of adding more links can be justifi ed. This the connectivity and distance between is divided into two stages. First, the ant model minimizes the structural changes web pages. The cost of two web pages is colony based model (Lin and Tseng, to the web link structure of a website and taken as a product of their distance and 2014) is applied to reorganize the link- also reduces the information overload to connectivity, so the most effi cient structure structure and then this resultant graph users. The model can be used to optimize is the one with the least possible overall structure is used in the second stage. In very large sized websites because this cost. the second stage, the local search (Singh model operates with the database of user Yin and Guo (2013) have developed a and Kaur, 2014) procedure is applied sessions to identify the number of links metaheuristic based on Tabu search (Yin to further improve the solution. This to be added and it does not remove the and Guo, 2013) approach for obtaining method takes more time than the Ant links which are already present in the web an optimal website structure. This Colony based procedure but it certainly structure. method also models Graph Structure generates a better link structure than Metaheuristic Techniques improvement problem as quadratic the Ant Colony based method. The local Metaheuristic methods have been found assignment problem type problem. The search procedure improves the solution to be very eff ective in improving the model is developed with the objective of step by step by removing the links with graph structure of websites with in less making the frequently used paths shorter. low weight and adding links with higher time. This section briefl y discusses such A 0-1 binary matrix represents a candidate weights.
CSI Communications | December 2014 | 13 Conclusions eff ective user navigation through web e-supermarket website,” IEEE Transactions: In this paper, a newly emerged problem of site structure improvement,” IEEE Systems, Man and Cybernetics – Part A, 36, website graph structure improvement has Transactions on Knowledge and Data 338–355, 2006. been discussed. Few main algorithms for Engineering PP (99) 1–18, 2013. [7] P Yin, Y Guo, “Optimization of multi- [2] M Dorigo, V Maniezzo, and A Colorni, criteria website structure based on solving this problem and improving the “The ant system: Optimization by a colony enhanced tabu search and web usage navigation effi ciency have also been briefl y of cooperating agents,” IEEE Transactions mining,” Journal of Applied Mathematics explained along with future research on System, Man, and Cybernetics, Part B, and Computation , 219, 11082-11095, directions. Optimizing the structure with 26, 1–13, 1996. 2013. the metaheuristic based approaches takes [3] C C Lin, “Optimal web site reorganization [8] Y Fu, MY Shih, M Creado, and C Ju, less time and can improve the large sized considering information overload and “Reorganizing web sites based on user structures. Mathematical programming search depth,” European Journal of access patterns,” International Journal model which works with web user sessions Operational Research, 173, 839–848, of Intelligent Systems on Accounting, 2006. Finance and Management. 11, 39–53, also seems a very promising method with [4] C C Lin, and L C Tseng, “Website 2002. very large sized website link structures. reorganization using an ant colony [9] H Singh and P Kaur, "Website Structure A lot of experiments are needed to be system,” Expert Systems with Optimization Model Based on Ant Colony performed with real and large web log Applications, 37, 7598–7605, 2010. System and Local Search", IJITCS, vol.6, data sets to discover more issues and [5] H Q Saremi, B Abedin, and A M Kermani, no.11, pp.48-53, 2014. DOI: 10.5815/ problems related to the above mentioned “Website structure improvement: ijitcs.2014.11.07 approaches. Very less amount of research quadratic assignment problem approach [10] E M Loiola et al. “A survey for the quadratic work has been done in this area and it and ant colony meta-heuristic technique,” assignment problem,” European Journal of Applied Mathematics and Computation, Operational Research, 176 (2), 657–690, off ers a tremendous scope for research. 195, 285–298, 2008. 2007. References [6] Y Wang, D Wang, and W Ip, n [1] M Chen, and Y U Ryu, “Facilitating “Optimal design of link structure for
Harpreet Singh is an Assistant Professor in the department of Computer Science & Engg. at DAV University Jalandhar. He is also pursuing Ph.D. from Guru Nanak Dev University Amritsar in the fi eld of Software Engineering. His research interests include Data Mining and Web Engineering.
Parminder Kaur is an Assistant Professor in the department of Computer Science & Engg. at Guru Nanak Dev University Amritsar, India. She completed her Ph.D. from Guru Nanak Dev University Amritsar in the year 2011. Her research interests include Component-based Software Engineering, Web Engineering and Software Security. About the Authors
CSI Communications | December 2014 | 14 www.csi-india.org Technical B. Raj Kumar* and M. Chandrakumar Peter** *Pre-fi nal year, Software Engineering, Periyar Maniammai University, Vallam, Thanjavur Trends **Assistant Professor, Software Engineering , Periyar Maniammai University, Vallam, Thanjavur Algorithm – AIM (Analysis in Minutes) Algorithm is an interpretable, fi nite set of instructions for dealing Sorting with contingencies and accompanying task that has recognizable Arranging elements in an order either ascending or descending is end-points, end-state or result for inputs given. It is a tool for called as Sorting . solving problems related to computational design. An algorithm is said to be correct if for every correct input, it halts with the Algorithm Time Complexity Worst correct output. Algorithms often have steps that repeat or require Case Auxiliary decisions until the task is completed. Diff erent algorithms can be Space written for the same task, i.e., using diff erent set of instructions, Complexity take more or less time, space or eff ort. Analyzing an algorithm Best Average Worst Worst means the resources that an algorithm needs. Resources can be memory requirements, communication bandwidth, logic gates Bubble Sort O(n) O(n^2) O(n^2) O(1) computation time etc. Analysis consists of two phases Bucket Sort O(n+k) O(n+k) O(n^2) O(nk) (i)Priori analysis Heap sort O(n log(n)) O(n log(n)) O(n log(n)) O(1) The bounds of algorithm computing time are obtained by formulating a function based on theory. It is independent Insertion Sort O(n) O(n^2) O(n^2) O(1) of programming languages and machines structures. The Merge sort O(n log(n)) O(n log(n)) O(n log(n)) O(n) stress is laid on the frequency of execution of statements Quick sort O(n log(n)) O(n log(n)) O(n^2) O(n) (ii)Posteriori analysis The actual amount of space and time taken by the Radix Sort O(nk) O(nk) O(nk) O(n+k) algorithms are recorded during execution. It is dependent Select Sort O(n^2) O(n^2) O(n^2) O(1) on the programming languages used and machine structures Searching Notation for asymptotic growth A search is an algorithm for fi nding an item among a collection of items. letter bound growth (theta) Θ upper and lower, tight equal Algorithm Complexity with respect Complexity (big-oh) Ο upper, tightness unknown less than or equal to Time with respect to ο (small-oh) upper, not tight less than Space (big omega) Ω lower, tightness unknown greater than or equal Average Worst Worst ω (small omega) lower, not tight greater than Binary search O(log(n)) O(log(n)) O(1) • Big O is the upper bound. Upper bound is specifi ed using Breadth First Search (BFS) - O(|E| + |V|) O(|V|) the notation Omega. Big O as well as Omega. Is represented using Theta. Representation of Theta can also be called as tight bound (it must be both the upper and lower bound). For Depth First Search (DFS) - O(|E| + |V|) O(|V|) example, an algorithm taking Omega (n log n) takes at least n log n time but has no upper limit. An algorithm taking Theta Linear (Brute Force) O(n) O(n) O(1) (n log n) is far preferential since it takes AT LEAST n log n Shortest path by O(|V||E|) O(|V||E|) O(|V|) (Omega n log n) and NO MORE THAN n log n (Big O n log n). Bellman-Ford • f(x) = Θ (g(n)) means f (the running time of the algorithm) Shortest path by Dijkstra, O((|V| + O((|V| + O(|V|) grows exactly like g when n. When input gets larger, the using a Min-heap as |E|) log |V|) |E|) log |V|) growth rate also gets higher (i.e asymptotically proportional priority queue to g(n)). • Same thing. Here the growth rate is no faster than g (n). Big- Shortest path by Dijkstra, O(|V|^2) O(|V|^2) O(|V|) oh is the most useful because represents the worst-case using an unsorted array behavior. as priority queue In short, if algorithm is __ then its performance is __
algorithm performance o(n) < n Conclusion O(n) ≤ n The above sections cover the time and space complexities of Θ(n) = n various searching and sorting algorithms used in Computer Science. Knowing the best, average, and worst case complexities Ω(n) ≥ n for various algorithms will help designers to design the data ω(n) > n storage and accessing system to the best possible combination. n
CSI Communications | December 2014 | 15 Technical Prof (Dr.) D G Jha* and Ms. Kimaya Ambekar** *Professor & Area Chairperson – IT; Programme Coordinator – MCA, K J Somaiya Institute of Management Studies and Research (SIMSR), Vidynagar, Vidyavihar, Mumbai Trends **Academic Associate- IT, K J Somaiya Institute of Management Studies and Research, Vidyavihar, Mumbai Algorithmic Computing: A Detailed Oriented Thinking about Procedures
Introduction 1. Start The exponential changes in the science of numbers and the use 2. Assume
CSI Communications | December 2014 | 16 www.csi-india.org the variable is positive, negative or zero 3. Output SV Algorithm: 4. Add SV to SUM (SUM = SUM + SV) 1. Start 5. Increase SV by 1 (SV = SV + 1) 2. Input
CSI Communications | December 2014 | 17 Recno() CustCode CustName CustType CustBirthDet The INDEX command creates a fi le CustCodeINDEX with referential Recno() i.e., 1 C102 Anil C 2 Jan 5, 2, 4, 1, 3 2 C101 Juhi B 5 Aug For retrieval the source fi le (referred to as underlying base data 3 C104 Manoj A 7 Dec fi le) along with it Index needs to be specifi ed; SAY 4 C105 Nimi B 8 Sept USE CustData INDEX CustCodeINDEX 5 C103 Onkar A 6 Feb The retrieved record will be in the order of CustCode (observe the Recno() column) (C103) is 4,3,5 in the three diff erent data fi les. • It causes redundancy as the data itself gets copied onto Recno() CustCode CustName CustType CustBirthDet destination fi le. 5 C101 Juhi B 5 Aug • Each of these data fi le are separate entities and 2 C102 Anil C 2 Jan any manipulation (such as append, edit and delete) performed one data fi le does not get refl ected on other 4 C103 Onkar A 6 Feb related data fi les causing inconsistency. 1 C104 Manoj A 7 Dec • The minimum disk space required for sorting process is 3 C105 Nimi B 8 Sept space equivalent to the source fi le i.e., in case, the disk space equivalent to the source fi le is not available, the sorting process cannot take place. Observation: • All these get corrected in INDEX concept. • The record number generated while entering the record in the data fi le remains’ the same in the index fi le, only THE CONCEPT OF INDEX the order of retrieval changes. Assume a data fi le that keeps the following details about the • Every time an INDEX command gets used only customers of a retail stores: referencing changes, the underlying base table remains the same. Recno() CustCode CustName CustType CustBirthDet • Since, all the manipulation (Append, Edit, Delete) 1 C104 Manoj A 7 Dec happens using underlying base table it minimises the data redundancy. 2 C102 Anil C 2 Jan • For index updation the command REINDEX update the 3 C105 Nimi B 8 Sept referencing INDEX fi le. 4 C103 Onkar A 6 Feb Another Instance: 5 C101 Juhi B 5 Aug USE CustData SORT on CustName to CustNameINDEX Name of Data File: CustData (original source fi le) The INDEX command creates a fi le CustNameINDEX with Note: A DBMS application creates a column Recno() by referential Recno() i.e., default and increments it by 1 every time a record is added. Whenever ordering of data is done using an INDEX option, 2, 5,1, 3, 4 it creates a new destination index fi le. Here the value of column For retrieval the source fi le (referred to as underlying base Recno() gets eff ectively used data fi le) along with it Index needs to be specifi ed; SAY For instance: An INDEX option can be specifi ed as- USE CustData INDEX CustNameINDEX USE CustData INDEX on CustCode to CustCodeINDEX The retrieved record will be in the order of CustCode (observe the Recno() column) The Working: The source data fi le CustData is: Recno() CustCode CustName CustType CustBirthDet Recno() CustCode CustName CustType CustBirthDet 2 C102 Anil C 2 Jan 1 C104 Manoj A 7 Dec 2 C102 Anil C 2 Jan 5 C101 Juhi B 5 Aug 3 C105 Nimi B 8 Sept 1 C104 Manoj A 7 Dec 4 C103 Onkar A 6 Feb 3 C105 Nimi B 8 Sept 5 C101 Juhi B 5 Aug 4 C103 Onkar A 6 Feb
CSI Communications | December 2014 | 18 www.csi-india.org Sort Algorithm Index Algorithm
1. Start 1. Start 2. Read degree and cardinality of the data from source fi le 2. Read degree and cardinality of the data from source fi le where degree= no of columns and cardinality = no of rows where degree= no of columns and cardinality = no of rows 3. Read data from source fi le in
The algorithms indicated here are to demonstrate the aspects of classifying and arranging data concepts using SORT and INDEX as discussed in the above case.
Evolution of Computing Paradigms The main mantra associated with any computing technique is creating applications that enable an increase in effi ciency of a business process. It makes sense here to briefl y look at various computing paradigms/disrupting technologies[5] that has seen exponential changes in the fi eld of computing and data processing. Features, pluses (+) and improvement scopes (-) of various disruptive computing models are listed below:
Personalised computing Smarter and personalized applications that adapt according to users behaviour and preferences are now being designed and developed.[24] The personalised computing model act as a tool with easy-to-use interface that enable managers to create environment that suits them the best. The following features characterises personalised computing: • Installation and maintenance of software is managed locally (+) • Customisable according to users’ need and specifi cation(+) • Utilisation is very low (-) • Exponentially very high up-front cost(-)
Reconfi gurable computing John Villasenor and William H. Mangione-Smith explains’ confi gurable computers as “Computers that modify their hardware circuits as they operate (are) opening a new era in computer design. Because they can fi lter data rapidly, they excel at pattern recognition, image processing and encryption”[7] The following features characterises reconfi gurable computing:[5] • Field Programmable Gate Arrays (FPGA) forms the main component (+) • Hardware is reprogrammable(+) • Programming time is slower (-) • Power consumption is invariably high (-)
CSI Communications | December 2014 | 19 Autonomic Computing Autonomic computing, a concept introduced by IBM in 2001, refers to IT systems being able to manage themselves dynamically adapting to changes in the computing environment, business policies and objectives. Autonomic functionality is characterized by four key areas, self-confi guring, self-healing, self-optimizing, and self-protecting.[6] The following features characterises autonomic computing:[5] • Inspired by Human Autonomic Nervous System (ANS) (+) • Motivation : rapidly growing complexities of integrating, managing and operating computer systems (+) • Increased complexity (-) Mobile Computing James Bucki states “Mobile computing is a generic term used to refer to a variety of devices that allow people to access data and information from where ever they are. It uses cell phone (apart from PC) connections to make phone calls as well as connecting to the Internet”.[2] It is widely regarded as technology on move since 1990s. The following features characterises mobile computing:[5] • No physical connectivity is required (+) • Intermittent (discontinuous) connectivity (-) • Limited bandwidth (-) • Dependent on mobile device maturity (-) Utility Computing Marios Alexandrou in online-article states “Utility computing refers to the ability of companies to access computing services, business processes, and applications from a utility-like service over a network. [The idea is to charge organisations as per the service that they use]. [While] the company off ering utility computing services can benefi t from economies of scale by using the same infrastructure to service multiple clients”[1] The concept of utility computing was conceptualised in 1960s; however it could materialised only in 1990s. The following features characterises utility computing:[5] • Provide community service for computing resources to business (processing power, bandwidth, data storage and enterprise software services) (+) Distributive Computing A distributed system uses software to coordinate tasks that are performed on multiple computers simultaneously. The computers interact to achieve a common goal, and they interact by sending each other an apt message. It is used to solve complex computational problems that cannot be completed within a reasonable amount of time on a single computer. The time necessary to complete all the calculations is reduced by harnessing the power of multiple computers.[3] The following features characterises distributive computing:[5] • Multiple autonomous computers connected through a communication network (though large complex programs are processed by multiple computers, distributive computing gives the illusion of one system) (+) • Provides transparency of resources (+) • The system has distributed memory where each processes gets its private memory (+) • Information is exchanged using communication models such as memory passing interface (MPI) (+) Virtualisation Virtualization refers to technologies designed to provide a layer of abstraction between computer hardware systems and the software running on them.[8] It enables: • Creation of virtual version of things such as an O/S, a server, a storage device or a network resource (+). For instance, logical partitioning of hard discs drive to create in eff ect two separate hard disc drives is a virtualisation of storage device Cloud Computing A model for delivering information technology services in which resources are retrieved from the internet through web-based tools and applications, rather than a direct connection to a server is referred to as cloud computing. Data and software packages are stored in servers. However, cloud computing structure allows access to information as long as an electronic device has access to the web. This type of system allows employees to work remotely.[4]
Conclusion and work backwards. Also, the ever evolving computing The concept of developing and designing The basic algorithmic structure techniques that defi nes the current algorithms forms the basis of any provides the clarity about the algorithmic generation digital fi rms is identifi ed for application creation. Every computing concepts, while the case presents the two future discussion. task requires step-by-step instructions in alternative logics that can be used for the References a logical sequence, and algorithm makes purpose of arranging and retrieving the [1] Alexendrou, M, nd. Utility Computing it easier to comprehend the requirements collected data using SORT and INDEX. Defi nition. Available at http://
CSI Communications | December 2014 | 20 www.csi-india.org infolifi c.com/technology/defi nitions/ html#lesson [Accessed 11 July 2014]. [7] Villasenor, J & Mangione- computer-dictionary/utility- [4] Investopedia., nd. Cloud Computing. Smith W H, nd. Confi gurable computing [Accessed 10 July 2014]. Available at http://www. Computing. Available at http:// [2] Bucki, J, nd. Defi nition of Mobile investopedia.com/terms/c/cloud- helios.informatik.uni-kl.de/ Computing. Available at http:// computing.asp [Accessed 12 July reconfigurable_computing/ operationstech.about.com/od/ 2014]. villasenor/0697villasenor.html glossary/g/Definition-Of-Mobile- [5] Sakr, M J, 2010. Introduction to cloud [Accessed 8 July 2014]. Computing.htm [Accessed 9 July computing. Available at http://www. [8] Waters, J K, 2007. Virtualization 2014]. qatar.cmu.edu/~msakr/15319-s10/ Defi nition and Solutions: Virtualization [3] Education Portal., nd. What is lectures/lecture01.pdf [Accessed 7 topics covering defi nition, objectives, Distributed Computing? - Principles, July 2014]. systems and solutions. Available at Environments & Applications. [6] Tivoli Software., nd. IBM developer www.cio.com/article/2439494/ Available at http://education-portal. Works : Tivoli autonomic computing. virtualization/virtualization- com/academy/lesson/what-is- Available at www.ibm.com/ definition-and-solutions.html distributed-computing-principles- developerworks/tivoli/autonomic. [Accessed 12 July 2014]. environments-applications. html [Accessed 8 July 2014]. n
Prof (Dr.) D G Jha is currently working as Professor and Area Chairperson - IT at K J Somaiya Institute of Management Studies and Research. He has over 25 years of experience and has authored a text book in the area of computing concepts and Management Information System. He is a Ph.D from University of Mumbai. He is also the programme coordinator of MCA. His area of interests are computing concepts, DBMS, Information systems, and HRIS. Ms. Kimaya Ambekar is Academic Associate- IT at K J Somaiya Institute of Management Studies and Research, Vidyavihar, Mumbai. She is a Masters in Computer Applications graduate from University of Mumbai and has two years of academic experience. Her areas of interests are Cloud computing, security, C# framework. About the Authors
CSI Communications | December 2014 | 21 Technical Ms. Seema Rawat*, Mr. Praveen Kumar**, Prof. Sunil Kumar Khatri*** and Dr. Balvinder Shukla**** *Assistant Professor, Amity School of Engineering & Technology **Assistant Professor, Amity School of Engineering & Technology Trends ***Director, Amity Institute of Information Technology, Amity University, Noida ****Vice Chancellor, Amity University Noida Recommendation Engines – A Generic Architecture Abstract: Have you ever imagined what makes web portal giant so easily predict your choice and recommend you products that a majority of time you are bound to purchase/refer/pay attention to? Well, these suggestions come from highly advanced systems having their own intelligence that run on your online footprint as inputs and predict your next purchase even before you go online. In technical terms, these systems are referred to as “Recommendation Engines” and it’s a booming term in the domain of IT Algorithms. Many international organizations have fi led patents against their recommendation engine algorithms while many others are in the development pipeline.
Introduction and involve complex pattern matching on 2. Collaborative: This approach A Recommendation Engine, in actual a set of predefi ned parameters and they relies upon direct learning about defi nition can be referred to as a system become effi cient with the increase in the the user which may require that can run on clustered / non clustered size of the content being fed to them. active or passive participation. environment taking user online footprint Approaches Adopted Active methodology involves as one of its input set and generating a The Recommendation Engines that have directly probing the user about probable footprint for the user thereby been developed so far run primarily on the his preferences through feedbacks providing its users a prediction closer to following approaches: and responses based on what reality. the user has purchased / surfi ng Recommendation Engines require a experience. The ratings that large dataset and a fast computing system a user gives in most websites that can perform analytics on the same against products/items are within fraction of seconds. With the rise in not mere ratings but an option the Big Data technologies both the above to save his priorities against components required for recommendation his own profi le that the engine engine to execute at its full potential are maintains. Cookie sniffi ng is one now easily available in the tech market. of the most common parts of Big giants of the internet like Amazon, this and comes under the passive Flipkart, IMDB, Google and many others methodology. Thus this approach get a majority of their sales from such expects the user to collaborate highly intelligent engines. There have with the recommendation been many instances where matrimonial engine in creating a profi le of his sites start showing up in primary ads 1. Content Based: This approach relies preferences. Popular organizations section of whatever one surfs online. on creating a plethora of parameters that use such an approach are These suggestions are a result of someone to describe a product ‘P’. Considering Pandora, MySpace and Facebook. accessing such sites – uploading/creating a smart phone as an example the 3. Hybrid Filtering: The collaborative profi les or showing interest in them. possible parameters could be screen For example, the DVD rental provider, size, image quality, Wi-Fi protocols, approach fails when the Netfl ix displays predicted ratings for every brand names, operating systems recommendation engine gets a displayed movie in order to help the user etc. The larger the parameter set “cold start” about the user – that decide which movie to rent. The online the better and easier it is to match is, when the user is very new and book retailer Amazon provides average patterns with user profi le and his his preferences are not available in user ratings for displayed books, and a online footprint. The parameters can the system to study and predict. list of other books that are bought by then be assigned weights and hence The best way in such case is to users who buy a specifi c book. Microsoft a relative priority is set for each of start creating the user footprint provides many free download for users, the parameter. All these parameters based on content based approach. such as bug fi xes, products and so forth. are then used to create a user But the content based approach When a user downloads some software, profi le and each time a prospective relies heavily on the parameters the system presents a list of additional user checks out another product, of the catalog items and is less items that are downloaded together. All his profi le gets updated. Hence we mature. In such scenarios it is these systems are typically categorized as see that the system learns about best suggested to use the Hybrid recommender systems, even though they the user preferences and selection Approach – Content based fi ltering provide diverse services. patters by his online footprint. for cold start like scenarios and Recommendation Engines, in simpler Popular platforms that use such an Collaborative for advanced / terms are programs that are data intensive approach are IMDB and Pandora. frequent users.
CSI Communications | December 2014 | 22 www.csi-india.org Algorithms feedback, storing user preferences, used. Some directly use a graph data The performance of a recommendation his past surfi ng records etc. The base for storing such information engine can be evaluated by testing how passive methods are more intrusive after the fi ltering has been done. well its collective intelligence behaves and include cookie sniffi ng, social 4. Pattern Matching: Once the profi le when fed with large data sets about engineering, and anonymous has been built the same is put to use user preferences, after all it is “machine tracking. The type and variety of the by the pattern matcher. Advanced learning” that the system relies heavily data could include clustering algorithms are used in upon. Data mining algorithms are a. Personal details about the user this phase. The pattern matcher used to generate the user profi le while such as age, location, favorite picks up quantum of information clustering algorithms are used to shortlist colors, nationality etc. from the profi le and starts building a the recommendations. The common b.Social details such as friends recommendation list. For example if algorithms include: of the user, family and the last browsing records of the user • n-Dimensional Euclidean Distance neighborhood details and indicates the user has purchased a • k-means algorithm how well the linking of their smart phone, the pattern matcher • Bayes Theorem preferences can be done with the would pick up the accessories that • Singular Value Decomposition preference information captured other user’s bought after such a • K-nearest Neighbor Algorithm about the user. purchase and would put this in the How a clustering algorithm works: c.Online footprint details such as recommendation engine basket with what the user surfs a majority an assigned priority. A Forrester of the times, what does he study on third party recommendation purchase, what is the average engines indicates that: spend of each of such purchase, a.15% of users admit to buying the prime categories of such recommended products. purchases etc. b.62% of the users that notice 2. Data Filtering: This process involves recommendations on websites removing or segregating the available follow them as it helps them fi nd data into information chunks. It a suitable product or accessory. processes an available data set and c. Vendors have also claimed that Assume that your system has all gives it a weight depending on how recommendation can increase the available parameters against a user relevant that information can be for the online sales by 2% to 20%. preference – everything that the user building the user profi le. For example, Limitation of Recommendation Engines has surfed/looked at so far. This could based on the location, age bucket, A recommendation engine in its core runs be single product cluster or a cluster nationality of the user we can measure primarily on abstract / un-confi rmed data of multiple of products. A clustering his online activity and link each of and hence is bound to make mistakes. algorithm simply fi nds the means from them to his profi le. If a user surfs all the Some of its limitations are: the un-clustered data (left graph) that product categories then age cannot 1. The algorithms run on static form of could be a single mean point or a set of be used as a determining parameter data – just because the user showed associated means. These can be found and hence should be assigned a lower interest on something today does not at the center of each of the cluster (right weight in this phase. Apart from the mean we can categorize and predict graph). From these means, the distance user specifi c information, this phase all his activity with respect to his to all the data points are calculated and also requires to engineer the available actions prevalent currently. Hence then based on which points are closest, social information such as the friend’s probability and trend analysis of such a small cluster is formed. This data set location, their age, their online activity engines needs to mature and take would indicate which products are viewed details into chunks of information this dynamic aspect of the data into more frequently by the user or are high in set for the next set of processes. account as well. the user’s surfi ng footprint. And hence we Sometimes what the friends of a 2. A profi le that matches another get a set of products that can be showed user are using (say a smart phone) profi le may not leave the same online to any user –whether the system has his could be closer to what the user footprint. Just because a set of users profi le maintained or for a fairly new user. would need next, if he does not own one. This phase performs best when in the given age bucket say ‘x’ from a Combined Architecture the available data is big, or in correct location, say ‘l’ liked a product does Hence a generic architecture of a terms – “Big Data” has been captured not mean a new user with the same recommendation engine, in terms of about the user. profi le may like one and is going processes involved can be summed up as: 3. Profi ling: Every recommendation to perform the same actions. The 1. Data Collection: This component engine has its own data structure of engines at present have no space to collects all the data about a user storing user information that in the accommodate exceptional cases and in an active or passive way. The end represents a user profi le. In a neither do they assume the existence active methods include taking user majority of engines, a graph structure of any such scenarios.
CSI Communications | December 2014 | 23 Conclusion algorithms at her disposal, and must make [2] Sucharita Mulpuru, Forrester Research,” In the past decade there has been a decision about the most appropriate What you need to know about Third Party a vast amount of research in the algorithm for her goals. Typically, such Recommendation Engines” http://www. fi eld of recommender systems, decisions are based on experiments, forrester.com/What+You+Need+To+Kno w+About+ThirdParty+Recomendation+E mostly focusing on designing new comparing the performance of a number ngines/fulltext/-/E-RES57914 algorithms for recommendations. of candidate recommenders. [3] Daniel Tunkelang ,”Recommendation as a An application designer who wishes References Conversation with the user”, 2011. to add a recommendation system to [1] Guy Shani,AselaGunawardana,”Evaluati n her application has a large variety of ng Recommendation Systems“, 2009.
Ms. Seema Rawat is working as Assistant Professor at Amity School of Engineering & Technology. She is M.Tech in Computer Science and B.Tech in Information Technology. She is pursuing her Ph.D in Computer Science from Swami Vivekanand University. She has 8.6 years of experience in academics. She has a number of international and national publications to her credit. She is a member of IEEE, IACSIT and IAENG. Her primary research area includes Cloud Computing, Data mining and Artifi cial Intelligence.
Mr. Praveen Kumar is working as Assistant Professor at Amity School of Engineering & Technology. He is M.Tech in Computer Science & Engineering. He is pursuing her Ph.D in Computer Science from NIMS University. He has more than 8 years of experience in academics. He has a number of international and national publications to his credit. He is a lifetime member of IETE, ACM, and IET. His primary research area includes Big Data Analytics, Cloud Computing and Data mining.
Prof. Sunil Kumar Khatri is Director in Amity Institute of Information Technology, Amity University, Noida. He has been conferred “IT Innovation & Excellence Award for Contribution in the fi eld of IT and Computer Science Education” in 2012 and “Exceptional Leadership and Dedication in Research” in the year 2009. He has edited four books, six special issues of international journals and published several papers in international and national journals and proceedings. His areas of research are Software Reliability, Data Mining and Warehousing and Network Security.”
Dr. Balvinder Shukla Vice Chancellor Amity University Noida is Ph.D from Queens University U.K. She is M.Tech from I.I.T kharagpur and has 27 years of rich experience in academia, industry, Research and Administration. To her credit she has presented many papers in seminars, conferences at national and international level. She has organized over 150 seminars, conferences and management developments programmes. She is on advisory board of many companies and is member of several professional bodies is closely
About the Authors associated with many CSR projects. Dr. Shukla is chair Person of more than 20 committees within Amity and outside.
CSI Communications | December 2014 | 24 www.csi-india.org Research Dr. Meghana Nasre Front Assistant Professor, Dept. of Computer Science and Engineering, Indian Institute of Technology Madras Stable Marriage – Algorithm and Variants
Introduction Stable Marriage The task of assigning agents to one another Our focus here is on the Gale and Shapley in 1962 introduced the or resources to agents is something that we algorithmic aspects of computing marriage problem where the participants routinely encounter in our day-to-day life. are n men and n women. Each participant For instance, consider the task of assigning such optimal allocations. ranks the members of the opposite gender students to colleges, or jobs to machines, by assigning a unique number from 1 to and roommates to one another. In many n. This ranking of the members of the At a very high level, the problems of these cases, participants are allowed to opposite side by a participant is called his/ of matchings under preferences can be specify preferences over other participating her preference list. If a man m assigns rank classifi ed into three broad categories: agents or resources. The goal then is to i to woman w and rank j to woman w (i) Stable marriage problem: here 1 2 compute an assignment that is optimal where i < j, we say that m prefers w to agents are usually referred to as 1 with respect to these preferences. These w . Consider any pairing M and let there be men and women and both men 2 problems are motivated by important a man m and a woman w not paired with and women specify preferences real world applications like the National each other in M. In addition, if both m and w over the members of the opposite Residency Matching Program (NRMP)[1], prefer each other to their current partners gender. the Canadian Resident Matching Service in M in the pairing, it is very likely that they (ii) Stable room-mates problem: here (CaRMS)[2], and NetFlix DVD rental will break their current engagements and the agents are of a single type program[3], to name a few. The scale of the pair off with each other. Such pairs are and every agent has a preference participants involved in these applications undesirable for the stability of a pairing. ordering of a subset of the agents. necessitates effi cient algorithms that Thus, the goal is to come up with a pairing (iii) House allocation problem: here compute optimal matchings respecting the of men and women such that there is no agents on one side (say people) specifi ed preferences. unstable pair. Formally, a pair (m,w) is said specify preferences on objects Problems of this kind are broadly to be unstable with respect to a pairing M belonging to another set (say classifi ed under the umbrella of if both m and w prefer each other to their houses). The houses in turn have “matchings under preferences”. Here the current partners in M. no preferences and therefore term matching is derived from the fact To illustrate these defi nitions, we these kind of problems are that the assignment can be viewed as consider a toy instance of the stable termed as problems with one- a matching in a suitably defi ned graph. marriage problem with three men {m , m , m }, sided preferences as opposed to 1 2 3 The history of this area dates back to and three women {w , w , w }. The the stable marriage which has 1 2 3 the seminal paper by Gale and Shapley[4] preference lists are as shown in Fig. 1 and two-sided preference lists. titled “College Admissions and Stability of can be read as follows: m1 prefers w to It is convenient to model the problem 1 marriage”, in which they defi ned the stable w and in turn m prefers w to w , and so as a graph G = (V, E) where vertices 2 1 2 3 marriage problem and gave a very elegant on. The instance admits several maximum correspond to the participating agents or algorithm to compute a stable marriage. matchings – the following two matchings objects and there exists an edge between Since its introduction, several researchers M = {(m , w ), (m , w ), (m , w )} and two vertices if both of them are mutually 1 1 2 2 3 3 from diverse domains like computer M' = {(m , w ), (m , w ), (m ,w )} are acceptable to each other. In addition, 1 3 2 2 3 1 science, economics, and game theory stable for the instance. It is easy to verify the preferences can be conveniently have developed a huge body of literature that the matching M˜ = {(m , w ), (m , represented by weights on the edges. The 1 1 2 with some deep and beautiful theory. w ),(m , w )} is unstable since the pair (m , assignment problem then, is to compute 3 3 2 2 A crowning glory to the eff orts of these w ) is an unstable pair with respect to M . a matching M in G which satisfi es the 2 researchers is the award of the 2012 Nobel desired notion of optimality. A matching in m : w , w , w w : m , m , m Prize for Economics to two eminent game 1 1 2 3 1 3 2 1 a graph is a set of edges such that no two m : w , w , w w : m , m , m theorists Alvin E. Roth and Lloyd S. Shapley. 2 2 3 1 2 2 1 3 edges share an end point. It can be easily m : w , w , w w : m , m , m The citation of their award mentions 3 3 1 2 3 1 3 2 observed that the graph derived from an that the prize is awarded for ‘the theory instance of the stable marriage problem or Fig. 1: A stable marriage instance of stable allocations and the practice of the house allocation problem is in fact a market design’. Our article attempts to bipartite graph. present the celebrated Gale and Shapley Given an instance of the stable marriage problem, it is not immediately stable marriage algorithm and also present A matching in a graph is a set of some optimality notions in a variant called clear that a stable matching always exists. the one-sided preference list model. Our edges such that no two edges However, Gale and Shapley proved that focus here is on the algorithmic aspects of share an end point. every instance of the stable marriage computing such optimal allocations. problem with strict and complete
CSI Communications | December 2014 | 25 preference lists admits a stable matching can be run when women propose men. possible allocation or a matching in the and also gave an algorithm that computes Such an algorithm outputs a women instance is M = {(a1, h 1); (a2, h 2)} which 2 one such matching in O(n ) time. Recall optimal stable matching; in the example leaves a3 unassigned to any house. Such that n denotes the number of men or the instance matching M' is indeed a women an allocation is clearly not acceptable if number of women in the instance. optimal stable matching. In fact, there is the size of the matching is only of concern. The algorithm: The Gale and Shapley an underlying lattice structure for the set This is because, there exists matching algorithm (shown as Algorithm II.1) is of all stable matchings for an instance M' = {(a1, h1); (a2, h3); (a3, h 2)} iterative and involves a series of proposals which has been exploited to generate which has larger size than the size of M. by one side, say the men. To begin with, stable matchings with additional criteria. However, we note that assigning a house [5] all men and women are unengaged. In We refer the reader to for a detailed to a3 is indeed at the cost of demoting every iteration, an unengaged man m exposition. a2 to a less preferred house. This brings proposes to the most preferred woman Numerous generalizations of the us to the definition of pareto-optimality w to whom he has not yet proposed. The stable marriage problem motivated from which is the weakest desirable property woman w accepts the proposal from m if real world applications have been studied. that any matching should satisfy. More either w is unengaged or w is engaged to These include allowing incomplete formally, we say that a matching M is a m' and w prefers m to m'. In the latter preferences, ties in preferences, hospitals pareto-optimal if no group of agents case, w breaks her engagement with m' resident problems, strategic issues related can improve their allocation without and gets engaged to m. It can be shown to the stable marriage, and more recently hurting someone. In our example, both that this proposal algorithm halts in O(n2) the study of stable matchings under social M and M' are pareto-optimal however, iterations, after which all men and women stability. This list is by no means exhaustive the matching M' is also maximum are engaged and the resulting matching is and pointers to detailed treatment of cardinality pareto-optimal. in fact stable. many of them can be found in[5],[6]. We
Algorithm II.1 Stable marriage algorithm ....a matching M is pareto- optimal if no group of agents can 1: Set all men and women as unengaged. improve their allocation without 2: while there exists an unengaged man m do hurting someone. 3: m proposes to the most preferred woman w to whom he has not yet proposed. 4: w accepts if either she is unengaged or she is engaged to m’ and w prefers m to m’. In fact, stronger notions of optimality 5: end while like rankmaximality, fairness, and popularity are studied. We discuss these The Gale and Shapley algorithm now move on to a restricted version of below. is deceptively simple and it is almost the stable marriage problem – that is, the Rank-maximality surprising that it terminates to output question of fi nding “good” matchings in A matching is rank-maximal if it matches a stable matching. Although we do not the house allocation problem. maximum number of agents to their [4] delve into the proof of correctness (refer One Sided Preferences rank-1 houses, subject to this, maximum for details), we note some interesting The one-sided model is motivated by number of agents to their rank-2 houses, properties of the algorithm and the several important applications like and so on. This very natural criteria was matching that it outputs. assigning graduates to training programs, introduced and studied by Irving[7] under Properties of the stable matching: families to government owned housing, the name of greedy matchings. A rank- First, observe that the algorithm does and mail based DVD rental systems like maximal matching can be computed by not specify the order in which men are NetFlix. In this model, members of only one assigning suitable weights to the edges considered in the while loop. Furthermore, side say the agents, specify preferences and transforming the problem into a as seen in the example instance, there over objects, say houses. Unlike the two- maximum weight bipartite matching can be multiple stable matchings in a sided model, where stability is accepted problem. However, this approach has the given instance. Yet, the men-proposing as a desirable notion of optimality, there issue of assigning large weights on edges stable marriage algorithm always outputs is no clear consensus about an optimal and thus the resulting algorithm has to a unique matching which is called the matching in the one-sided world. men optimal stable matching. In a men Fig. 2 shows an instance of the house optimal stable matching, every man m allocation problem with 3 agents and 3 A rank-maximal matching can is guaranteed to be assigned to the best houses. A possible women amongst all the women be computed by assigning he can get in diff erent stable matchings suitable weights to the edges a1 : h1, h2, h3 of the instance. In the above example, M is a2 : h1, h2, h3 and transforming the problem a men optimal stable matching. Since the a : h1, h2 3 into a maximum weight bipartite two sides men and women are symmetric, it is natural to expect that the algorithm Fig. 2: A house allocati on instance matching problem.
CSI Communications | December 2014 | 26 www.csi-india.org deal with arithmetic on large numbers. instance. An agent wishes to switch Discussion A simple combinatorial algorithm from matching M to another matching We have presented here a bird’s eye-view of was presented by Irving et al.[8] which M' if he gets a better preferred house the stable marriage problem and some of its avoids this issue of arithmetic with large in M'. Popular matchings have another variants related to matchings in one sided numbers. The efficiency of the algorithm attractive feature that they rely only preference model. A topic as rich as this and stems from an elegant use of the Gallai on the relative ranks of the houses of immense practical importance has been Edmonds decomposition which is well- as opposed to rank-maximal and fair dealt in detail in several research papers and known in matching theory. matchings which depend on absolute books two of which we cite here[5],[6]. Both Fairness ranks. This makes popularity useful even of these books provide lucid and accessible Rank-maximal matchings aim to assign when the absolute preferences of agents treatment of intriguing issues related to as many agents as possible to their highly are not known or cannot be revealed. the stable marriage problem and also point preferred houses. Fair matchings work Although natural and appealing, popular out important open problems in this area. the other way around by assigning least matchings have a downside that there are We hope that the quest to resolve some number of agents to their less preferred instances in which no popular matching of these will enhance our understanding houses. More formally, a matching M exists. In fact, the instance shown in of the beautiful structure underlying these is fair if it leaves the least number of Fig. 2 does not admit any popular problems. matching. On the positive side, when agents unmatched (and therefore is a References popular matchings exist, they can maximum cardinality matching), subject [1] A E Roth, “The evolution of the be considered stable since no group to this, the least number of agents are labor market for medical interns of agents can force migration to any assigned to their last choice houses, and residents: A case study in game other matching. Abraham et al.[9] and so on. Fair matchings can again be theory,” Journal of Political Economy, have presented efficient algorithms computed by assigning suitable weights vol. 92, pp. 991–1016, 1984. that determine if an instance admits a on edges and transforming the problem [2] “Canadian resident matching popular matching and computes one if into a minimum weight bipartite service.” it exists. matching problem. [3] “Netfl ix dvd rental.” As remarked earlier, there is [4] D Gale and L Shapley, “College no clear consensus on the notion of admissions and the stability of optimality in the one-sided world. It Fair matchings can again marriage,” American Mathematical is upto the application to chose the be computed by assigning Monthly, vol. 69, pp. 9–14, 1962. right notion as per their requirements. [5] D Gusfi eld and R W Irving, The Stable suitable weights on edges and It will be interesting to study how Marriage Problem: Structure and transforming the problem into these notions compare with respect Algorithms. MIT Press, 1989. to different parameters like average a minimum weight bipartite [6] D Manlove, Algorithmics of Matching rank of the agents, effect of length matching problem. under Preferences. World Scientifi c of preference lists, and on the prior Publishing Company, 2013. knowledge if any, about the distribution [7] R W Irving, “Greedy matchings,” from which preferences of agents are In the instance in Fig. 2, the Tech. Rep. TR-2003-136, University drawn. Findings of such a study (either matching M' is both rank-maximal as of Glasgow, 2003. experimental or theoretical) may serve well as fair for the instance. However, [8] R W Irving, T Kavitha, K Mehlhorn, as a guideline to chose a particular as expected, there are several instances D Michail, and K Paluch, “Rank- notion of optimality. where rank-maximal matchings and fair maximal matchings,” ACM matchings are different and one may be Transactions on Algorithms, vol. 2, no. desirable over the other. .... there is no clear consensus 4, pp. 602–610, 2006. [9] D J Abraham, R W Irving, Popularity on the notion of optimality in the Popularity is another appealing notion T Kavitha, and K Mehlhorn, “Popular of optimality which can be defined one-sided world. It is upto the matchings,” SIAM Journal on informally as follows. A matching M is application to chose the right Computing, vol. 37, no. 4, pp. 1030– popular if no majority of agents wants notion as per their requirements. 1045, 2007. to switch to any other matching in the n
Dr. Meghana Nasre is currently engaged as an Assistant Professor at the Department of Computer Science and Engineering of Indian Institute of Technology Madras, Chennai. She has received her Masters (M.Tech., CSE) from IIT Bombay and her doctorate in Computer Science from IISc Bangalore. Her research interests include graph algorithms, especially matching under preferences, and dynamic graph algorithms. About the Author
CSI Communications | December 2014 | 27 Research Srabani Mukhopadhyaya Front Associate Professor, Birla Institute of Technology, Mesra, Kolkata Campus Generating Random Numbers and their Applications in Computing
Randomization and Probabilistic problems (in scientifi c terms these are algorithm, how can it be random! Strictly techniques have signifi cant applications NP-complete, NP-hard). Some of these speaking, it is not possible to generate in almost every branch of modern science problems, though hard to compute for a random numbers using an algorithm. ranging from genetics, evolutionary set of pathological input instances, are However, it is possible to generate a biology to modern economics. We have actually not so diffi cult to compute for sequence of numbers which have the learned to accept randomness as an most of the cases. Probabilistic theory desirable properties of a random sequence. essential feature in modeling and analyzing analyzes and explains this phenomenon. Two most important properties among nature. In modern computer science also Actually, if the inputs are drawn randomly these are (i) generated numbers should probabilistic methods and randomness from a probability distribution on the set appear to be uniformly distributed over play a very important role in a wide range of all inputs, it is likely to get an instance the range [0, 1] and (ii) there should not of applications including combinatorial which is easy to solve. Instances which are be any correlation among these numbers. optimization, computational biology or hard to solve appear with relatively small Apart from these desirable properties even in communication networks too. For probability. For example, there are several of the generated random numbers, the the last few decades, we have witnessed NP-hard graph problems. The question random number generating algorithms a tremendous growth in the use of whether these problems are hard to solve are expected to possess some features randomness and probabilistic theory in for most of the instances or for a relatively from the application point of view. Some the fi eld of computing. small fraction of graphs can be answered of these can be listed as (i) the generators Randomness in Algorithmic Computing by using random graph models. must be time and space eff ective (as is the and Analysis Another reason behind popularity case for any algorithm), (ii) the generator There is a class of algorithms, known as of randomized algorithms is its effi ciency must be able to reproduce the identical randomized algorithms, which incorporate and simplicity. In many applications, stream of random numbers (this is not randomness in their execution. In actual randomized solutions are more effi cient only useful for debugging but also required practice, these algorithms use a random than the best known deterministic for comparison of two processes in case value, generated by random number solution. There are several NP-hard/ of simulation) (iii) the generator must be generator, to decide the next step at NP-complete problems, whose effi cient able to produce multiple sequences. diff erent stages of their execution. For randomized solutions can be designed Early Methods example, in randomized quicksort, at every at the cost of marginal degradation of In earlier days the random numbers were recursive step, the pivot element is selected accuracy. However, in these cases, the generated by hand, such as throwing dice, randomly to impose randomness in the randomized algorithm can be tuned in drawing cards from well-shuffl ed pack, input instances. Worst case complexity such a way that the probability of getting drawing numbered balls from stirred analysis of the quicksort shows that there an inaccurate answer can be made less are a few input instances for which quicksort than a prefi xed level. Primality testing is takes longer amount of time compared to one such problem. Strictly speaking, it is not its average execution time. Therefore, in Random Number Generators possible to generate random repeated applications of this algorithm, For implementation of randomized numbers using an algorithm. it is desirable to impose randomness in algorithms, generation of random numbers the input instances to prevent excessive is essential. Moreover, a simulation of any However, it is possible to occurrence of unfavorable instances. In process which involves an inherent random generate a sequence of numbers these kinds of randomized algorithms, component, make use of random numbers. which have the desirable the execution time varies with the set of In this article we are trying to indicate a properties of a random sequence. random values used in their execution few methods for drawing random values without aff ecting the result. from a uniform probability distribution Complexity theory classifi es some of over the range [0,1]. It may be noted that if urn, etc.[1]. Similar methods like picking the computational problems as diffi cult we can generate uniform random variates, up numbers 'randomly' from telephone it is possible to get samples from most of directory or from expansion of π, etc. are the common distributions using suitable still followed in informal situations. randomized algorithms .... use transformations. Mechanical and Electrical Devices a random value, generated by On fi rst thought the idea of an Later, in the early twentieth century, random number generator, to algorithm for generating random numbers when statisticians joined gamblers for decide the next step at diff erent sounds like an oxymoron. If something is their academic interest, some mechanical computed deterministically, as it would be stages of their execution. devices were designed for faster in the case of the output of a deterministic generation of random numbers. Some
CSI Communications | December 2014 | 28 www.csi-india.org of these devices are still found in many > k) then exactly the same sequence will variate. To name a few, chi-square test, lotteries. One such instrument was a be generated once again from that point serial tests, run test are well known, widely rapidly spinning disk, which was fi rst used (Xp) and this cycle (Xk, Xk+1, Xk+2, …, Xp) will used empirical tests. However, which of by Kendall and Babington-Smith in 1938[2]. be repeated endlessly. The length of this these tests are best to use is a diffi cult Later electric circuits based on randomly cycle is called the period. The maximum question to answer. It can only be said that pulsating vacuum tubes were used to period can be m and the corresponding the test should be selected carefully so that generate random numbers at much faster LCG is said to have a full period. It is always it is consistent with the use of the random rates. The Electronic Random Number desirable to have an LCG with a full period. numbers generated. Indicator Equipment (ERNIE) was used There are two variations of LCG, While we have seen a great deal of by British General Post Offi ce[3]. A recent mixed LCGs, when c > 0 and multiplicative applications of randomization in computing, device based on counting gamma rays is LCGs when c = 0[7]. one can safely predict that this is only the [4] described in a paper by Miyatake at. al . General Congruences beginning. More innovative applications in Midsquare Method A more general congruence can be the future would require pseudo random During 1940 to 1950, researchers started defi ned as numbers which are closer to the actual ones. It would lead to more sophisticated designing numeric or arithmetic methods Xi = f(Xi–1, Xi–2, . . . ) (mod m) and Ui = Xi /m; to generate random numbers. These algorithms as well as hardware techniques For a larger period, two variations of this for generation of random numbers. methods are essentially sequential and function f are each new random number is generated (1) quadratic congruential generator: as a function of its predecessors. First 2 f(Xi–1, Xi–2, . . . ) = aX i-1 + bXi-1 + c It can only be said that the test and the most famous such arithmetic (2) multiple recursive generator (MRG): should be selected carefully generator, proposed by Von Neumann, f(Xi–1, Xi–2, . . . ) = a1Xi-1 + a2Xi-2 + ... + aqXi-q was the midsquare method[5]. This method so that it is consistent with the Another widely used random use of the random numbers starts with a seed value X0 of four digits. number generators are feedback shift 2 In X0 the middle four digits are taken register generators. These generators generated. with a decimal point to the left of it as are developed on the basis of a paper X1. Following the same procedure X2 published in 1965[8]. References 2 is generated from X1 . This method is [1] Hull, T E and A R Dobell, “Random continued till the desired number of Number Generators,” SIAM, Rev., Vol. 4, random numbers are generated. ... widely used random number pp. 230-254, 1962. [2] Kendall, M G and B Babington-Smith, Linear Congruential Generators generators are feedback shift “Randomness and random Sampling The most widely used random number register generators. numbers,” J Roy Statist. Soc., Vol. 101(A), generators in recent years is Linear pp. 147-166, 1938. Congruential Generators (LCGs). This [3] Thomson, WE, “ERNIE – A Mathematical technique was introduced by Lehmer in Testing Random Number Generators and Statistical Analysis,” J Roy Statist. Soc., 1951[6]. This method also starts with a We have already mentioned that a Vol. 122(A), pp. 301-324, 1959. seed value X0. A successive sequence deterministic algorithm cannot generate [4] Miyatake, O, M Ichimura, Y Yoshizawa of numbers X1, X2, … are generated by a truly random numbers. For example, if and H Inoue, “Mathematical Analysis recursive formula we look at the formula used in LCGs, by of Random Number Generator Using mathematical induction it can be proved Gamma Rays,” Math. Jap., Vol. 28, Xi = (aXi–1 + c)(mod m) and Ui = Xi/m; c(ai–1) pp. 399-414, 1983. i = i, 2, 3, ... that for i = 1, 2, 3, ... X = (ai X + ) [5] Von Neumann, “Various Techniques Used in
i 0 a–1 where m, c, a and X0 are all non-negative v Connection with Random Digits,” Natl. Bur. integers and a, c, and X0 are all less than m. mod(m) Ui = Xi/m That is, the random Std. Appl. Math. Ser., Vol. 12, pp. 36-38, 1951. numbers can be generated deterministically [6] Lehmer, D H, “Mathematical Methods in Large Scale Computing Units, Ann. The most widely used random if the seed is known. These pseudo random numbers are generated with an Comput. Lab., Harvard Univ., Vol. 26, pp.141-146, 1951. number generators in recent expectation that they would appear to be years is Linear Congruential [7] Knuth, D E, The Art of Computer independently and identically distributed programming. Vol. 2., Seminumerical Generators (LCGs). over [0,1]. At this stage, these random Algorithms, Addison-Wesley, 1998. number generators would undergo several [8] Tausworthe, RC, “Random Numbers It is obvious from the formulation tests which will ensure how good the Generated by Linear Recurrence Modulo that Xi s can have values between 0 and generator is in a sense that a good random Two,” Math. Comput. Vol. 19, pp. 201-209, m-1. In the sequence, if a value appears number generator produce random values 1965. again (say, Xk and Xp are identical and p which represent a true IID U(0,1) random n
Srabani Mukhopadhyaya received M.Sc. in Applied Mathematics from the University of Calcutta in 1990. She received her Ph.D. in Computer Science from Indian Statistical Institute, Kolkata, in 1997. Currently, she is an Associate Professor at Birla Institute of Technology, Mesra, Kolkata Campus. Her current research interests include swarm intelligence, graph and combinatorial algorithms, parallel and distributed computing, sensor networks, etc. About the Author
CSI Communications | December 2014 | 29 Rajesh Sharma Article Director, Telecommunication Engineering Centre, New Delhi
Challenges in Using Aadhar as Unique Identity Number for Delivery of e-Government Services
The Unique Identifi cation Authority of to be addressed for obtaining maximum such capacity may result in frequent India (UIDAI), which functions under the benefi t of such a mammoth and expensive failure of the system as seen in the Planning Commission of India has been project. Some of these challenges are case of IRCTC in recent times. entrusted with the responsibility of issuing listed below along with recommended Security and privacy issues with the data: cards to the citizens of India. Aadhaar is steps that can be taken for removing such The data is being collected by the UID a 12-digit unique identifi cation number hurdles in technical, organizational and with the help of private agencies. There is issued by the Indian government to every legal domains: strong apprehension expressed by various individual resident of India. Storage and computational challenge: The quarters that they may misuse the data The Aadhaar project was initiated as project will require storage and processing and pass it on for commercial gains or for an attempt towards having a single, unique of huge amount of data. The scope of the tracking individuals. Moreover, there is identifi cation document or number that project roughly translates to capturing 12 threat of hacking of the data for snooping would capture all the details, including billion fi ngerprints, 1.2 billion photographs, as well as possibility of destroying the demographic and biometric information, and 2.4 billion iris scans. The fi le size for centralized data with cyber attacks that of every resident Indian individual. It was each enrolment is approximately 5 Mb. have become common in recent times. felt necessary because of the fact that For 1.2 billion people, the fi le size would Citizen forums have become active and currently there are a plethora of identity be measured in pentabytes. Moreover, are protesting against such possibilities. documents in India including passports, the captured data exhibits the three The possible solution to such permanent account numbers (PANs), characteristics of Big Data namely - Volume, problems are as follows: driving licenses and ration cards but Variety and Velocity. Updation of dynamic • Policy of data disclosure: Several there is no Single Identity Number (SIN) data of citizens such as address, occupation, countries have implemented that can be used for uniquely identifying marital status etc in a secure and immediate strict laws against misuse of such the citizens. Further, there was a vast way is another important challenge that sensitive personal information segment of citizens who did not have any needs to be addresses. In the absence of that includes obtaining permission identifi cation documents to claim the availability of such processes, the data of owner of data before using it, benefi ts of government schemes. It was stored may become old and unreliable. obtaining consent of citizens before targeted to provide 600 million such cards The solution to this important issue is divulging to third parties. to the citizens over a period of 4 years not a trivial task and may require following • Encryption of data: Austria has put from the year of inception, i.e., 2009. Over action on the part of UID: in place a system in 2004 in which time it will cover the entire population of • Establish a mechanism by which a source identifi cation number the country which is presently about 1200 citizens can log a request for change sPIN is generated from citizen data million but likely to increase further by the in dynamic information contents. for each sector. This encryption time the project nears completion. The The request should be routed mechanism ensures that use of cost of the project has been estimated for physical verifi cation in a time data outside the intended sector to be as high as Rs 45000 crores by its bound manner to the appropriate is impossible. Perhaps similar critics while the government maintains authority in district administration. mechanism may be put in place in that it may not cross Rs 18000 crores. This will require use of technology India for ensuring data security to It was envisaged that the project as well as suitable process for fi xing win trust of citizens. can be used for extending benefi ts of of responsibility with monetary • Securing of data against cyber such schemes directly by money transfer penalties to the concerned offi cers attacks: This can be achieved by using to the bank accounts of eligible citizens. (as has been implemented for time distributed data storage, securing the Another important application of Unique bound response to RTI applications) storage area physically as well as Identity Number is that it can be used • Ensure that scalability of the data logically, use of eff ective fi rewalls, as the primary key for linking databases centres and data warehouses that are restricting access to data and so on. maintained by various government designed for storage of the huge data. Organizational and Legal issues: There are departments such as Income Tax, Passport • Ensure that suffi cient computational several organizational issues as various Offi ces, Regional Transport Offi ces power is available at the data departments feel threatened by loss of etc. Seamless connectivity achieved by centres because the transactions their control over the processes which linking of databases with proper semantic with requirement of verifi cation of are presently in their domain. These standards will enable these departments to data will multiply as interoperability apprehensions relate not only to giving up share documents online and substantially features in e-government services power and less avenues for corruption but reduce the hassles of paperwork required get activated which is likely to be are also directed to prevent close monitoring for providing citizen services. However, mandated by the government in the of their activities which are possible once the there are several challenges that need new NeGP shortly. Failing to ensure systems interoperability is achieved by this
CSI Communications | December 2014 | 30 www.csi-india.org project. It is necessary to overcome these • Legal issues regarding use of the acceptance will take off dramatically. organizational and legal issues as follows: UID number in legal proceedings One such use that comes to mind is that • By making the public aware of the can be sorted out with proper of grievances reddresal. At present people benefi ts that will accrue to them by legislations by the government. complain for government services by letters, implementation of such projects. At present this issue is the most emails etc which are diffi cult to verify by The best way is to demonstrate critical that is being faced by the redressal authorities. If UID number is the benefi ts of direct transfer of this project as Supreme Court used, authenticity of such complaints can subsidies to them. This was tried by has prohibited government from be easily ascertained and acted upon. the government but was opposed making UID number mandatory. The progress on providing Aadhar by citizens as the UID number was Government has to clearly explain is expected to pick up momentum with made mandatory without provision to the courts how such a move will the decision of the government for of UID numbers to them. Instead adversely aff ect the e-government continuing the project. Since the unique of such arm twisting, it is better to interoperability capabilities for identity for all citizens is now a reality, keep both options open for some delivery of government services. the UIDAI must quickly move to address time and let people themselves Developing innovative applications for use the complex issues related to technology, judge the benefi t of UID. Once of the data collected: The fourth critical security, privacy, organizational and voice is raised from their side, challenge is to use the huge data that is legal complications in order to promote the resistance of government collected for some innovative applications its voluntary acceptance among the departments will eventually end instead of letting it lie for just a handful citizens and make Aadhar as a facilitator and UID will be adopted by the services. Once the innovative applications for eff ective delivery of e-Government citizens willingly. are useful to a large section of population, services. n
Rajesh Sharma Belongs to Indian Telecom Service (ITS), Ministry of Communication & IT and posted as Director at Telecommunication Engineering Centre, New Delhi. He has experience of more than 24 years in development and maintenance of telecommunication infrastructure in various capacities. Presently he is working on doctoral research at Indian Institute of Management, Indore in Information System area. His areas of interest include adoption of e-Government, interoperability capabilities, network security and business processes. The views expressed by the author are his own. Mobile number: 9425001221, Email: [email protected] About the Author
CSI Communications | December 2014 | 31 Nandakumar Edamana Article B.Sc. Computer Science student, College of Applied Science, Vattamkulam
Defi ning and Describing Multilayer Approach for Safe Social Networking
Abstract: Social media is becoming an important part of modern life. But most social networking systems (such as social networking websites) are threatening users' privacy, security and health. They are designed in a way to make the users addicts. They present a lot of features which are enabled by default, and make a big part of the data given by the user public without explicit notice. Moreover, many 'useless' features (i.e., features which cannot be used for any creative/reproductive work) of social media are wasting a lot of human resource and energy. But dividing the features of a social networking system into diff erent modules (in the user-level), which are enabled only on the user's request can make the system safe, private, and non-addictive.
Defi nition are two cases: The word 'necessary' should be taken in Simple: Generally speaking, Multilayer 1. The user likes to use certain features a programming point of view, otherwise Approach for Safe Social Networking only. But some features/notifi cations are the system owners can fool the public (MASS) is a social networking system that turned on automatically and some private by declaring everything as necessary and is safe, private and non-addictive. data is made public without the knowledge/ telling "everything will be useful for you in Real: The term 'Multilayer Social consent of the user. (S)he is annoyed by this social networking system." Networking' means dividing a social frequently received messages related with Forbidden Layers networking system (such as a social those features. These are truly server-side layers, which networking website) into diff erent 'global' 2. The user starts a social media the user cannot access directly. In MASS, and 'local' 'layers.' A 'layer' means a subset account just for fun. Being eager to API source fi les and user's private data are of the features provided by the system. A explore, (s)he starts to use all features kept in forbidden layers. Any upload from user of the system must go through global of that social networking system. In an a user should be kept in forbidden layers layers while the local ones can be skipped everything-enabled-by-default social until the user asks to make it public (many as (s)he wishes. Only necessary features networking system, the user can easily do contemporary website don't do this). For like authentication should be made global this, so there is no chance for hesitation/ example, in the case of a regular Apache so that the user gets maximum control to double-thinking. web server, the contents of forbidden turn off (or not to use) the features (s)he It is clear that case 1 is a major layers (including the uploads) are kept doesn't like. When there are global layers privacy/security issue, and case 2 is the outside the 'public_html' directory. It can for security, privacy, non-addictiveness cause for social media addiction and be external servers also (with similar and other ethical aspects in a Multilayer related health risks. The terms of service privacy and security policies). Social Networking system, the system is of most websites will be justifying these said to be a Multilayer Approach for Safe problems to make them legal, while they Social Networking (MASS) based system. remain unethical. Hence a new social Description networking strategy is needed in which the user gets maximum control over the Background features and which respects ethics. Social networking systems are becoming the primary mode of communication and Global and Local Layers meeting in this IT era. They can be used The idea towards the solution was to to organize people, share news and ideas, divide the features provided by the social and run socially important campaigns, networking system into subsets, so that irrespective of the location of the involved. the user can choose the subsets (s)he Fig. 1: General arrangement of a multi layer social networking system. Note that LL1 is not related They are preferred over traditional needs, without involving with the other with FL1, LL2 is not related with FL2, etc. communication methods because of their ones. However, while making this idea swift nature and fl exibility (even sound practical, developers would fi nd that and video can be sent easily, to thousands many API (Application Programming Nested Layers at a time). Interface) features like authentication A 'nested local layer' is a local layer which All major social networking websites should be run globally. So the idea arose comes under another local layer ('hosting come with a variety of features such as that the global features like authentication local layer'). Let GL1 be a global layer. Local groups, discussion, microblogging, instant can be made 'global layers' which are layers LL1 and LL2 come under this. Now a messaging, sharing, etc. (although the spread everywhere in the system, while user has to pass through GL1 in order to names can vary). Many users join a social all other features like involving in groups reach LL1 or LL2. Let NL1 and NL2 be two networking website to use only certain and uploading images (which are not local layers which come under LL2. Now features of it, but they soon start to use a necessary) should strictly be made 'local LL2 is called a 'hosting local layer' and NL1 lot of things they didn't really want. There layers' so that the user can control them. and NL2 are called 'nested local layers.'
CSI Communications | December 2014 | 32 www.csi-india.org Here a user has to pass through GL1 and second half is the presence of mechanisms cases). But features asking for the co- LL2 in order to reach NL1 or NL2. which prevent the leak of users' private data. operation of the service provider (such as Generally speaking, hosting a single A system can make its privacy-killing tricks precautions for non-addiction) cannot be layer under a local layer makes no sense legal by showing them in its privacy policy ignored at any excuse. since they can be combined as a single and forcing the user to agree that, but it Layer for Client-side Encryption and local layer. However, the former can be no longer becomes a MASS-based system Decryption applied when there is a chance for more since the term is related to ethics too. This can be a global layer or a hosting nested layers in the future. Privacy issues cannot be solved without the layer of local layers which handle private help of the server/service provider, hence informations of a user. Its availability is the privacy layer is a must in a MASS-based necessary. Messaging and private data system. storage features can make use of it. But Security layer should contain function it is sometimes dangerous, so the user which prevents a user from attacking the should get the freedom to skip this. server and other users. It also contains the This layer helps the user to encrypt act of maintaining fi le permissions, which the data with a password before sending keeps the users' data private. The user to the server. Being encrypted, the must be off ered with the option not to list service providers cannot read the data. If their data/profi le in search engines. This Fig. 2: A simple system having nested the data is a message, only the intended and related features can be combined into local layers recipient who knows the password can an 'anonymity layer,' if needed. decrypt it. Very simple algorithms are Actions against the server's privacy enough to do this job, given the encrypted Extended Defi nition are harmful to other users' privacy. Also, text can be decrypted only with the Although the term suggests covered the server shouldn't harm the users' data correct password. Client-side programs systems to have divided into layers without and client devices. Privacy layer should are usually written in JavaScript, so the linking to each other, MASS means more. also consider this 'reverse' issue. Hence source code is publicly-viewable. It is not The overall idea is introduced to protect the Security and Privacy layers are two- a problem since the password from the the user's privacy, tastes and health. way systems. user is (should be) unpredictable. Also, Multilayer Social Networking seems to be Some features like Secure Socket users can verify their data is not sent to implemented already to an extend in many Layer come under both security and the server before the encryption, if the services (we don't know what happens in privacy layers. So it is better to combine source code is available. their servers) although a clear defi nition these two layers into a single one. However, client-side encryption is was not available. But it is doubtless that Authentication (log in system in websites) not possible for the data which has to be Multilayer Social Networking alone can't also comes under this layer. read by the server. For example, user's provide safe social networking. a MASS- Irreversible techniques such as publicly viewable posts and uploads based system should always contain some hashing should be used to store the shouldn't be encrypted in this way. Also, 'ethical' global layers to do this. passwords. there is a risk of data loss if the password is forgotten. Global Layer for Security and Privacy Global Layer for Health Care Newsletter Service Privacy layer is a must and should Health layer consists of features or Newsletter services which follow MASS always be implemented as a global one. limitations which can prevent users from or which are included in a MASS-based 'Implementing a privacy layer' doesn't social media addiction and related health system should be compatible with the lonely mean the creation of some API risks. This layer may include features like: following: functions to ensure privacy, but the absence • Dark themes (visual styles) like 'High 1. The publisher can request anybody to of surveillance/'data mining' mechanisms Contrast Inverse' which off er less join the newsletter, but (s)he cannot also (creating secret profi les by watching a strain for the eyes. add people's id (username, email, ...) user's actions rather than by asking him/her • Alarms to prevent the user from directly to the recipients' list. questions is also surveillance). That means, being a social media addict. 2. The publisher cannot view the half of a privacy layer can be implemented • Limits on the number of likes/ recipients' list; otherwise (s)he can by simply removing the surveillance traps. comments that a user can put per take a copy of the list, run other In other words, a social networking system day. unsolicited newsletters, and send which doesn't steal/share users' private data • Readable text styles with optimum newsletters to unsubscribed people. has already implemented the privacy layer font size and spacing. 3. The user has to confi rm his/her partly. However, software daemons which Users' requests should be fi ltered using subscription in order to receive issues watch users' actions to detect terrorism is the health layer. However, the health layer and has the freedom to unsubscribe allowed. But data from them which are not can be made less strict since the user can at any time. related with terrorism shouldn't be recorded. take care of his/her heath by himself/ Detailed view is given under subtitle Privacy herself (security and privacy can only be Account Deletion and Data Deletion vs Safety. established with the help of the service The user has the freedom to delete the While the fi rst half of the privacy layer is provider; but health care does not require account whenever (s)he wants. Access the absence of surveillance mechanisms, the the co-operation of the a server in most to this feature should not be hidden
CSI Communications | December 2014 | 33 somewhere, but easily accessible. All data • No unsolicited newsletters and related with a particular service sould be notifi cations. deleted from the server/backup facilities Pros and Cons once the user removes himself/herself from it. In the case of account deletion, Pros everything related with the user should be 1. The user gets maximum security and removed. privacy. Removal of already shared data 2. Social media addiction can be will not be possible always. If so, a clear controlled to an extend. message about this should be shown in all 3. Human resource is not wasted and share dialogs. For example, a 'share this can be used for good purposes. For example, in a usual social media picture publicly' dialog shows a message 4. Eco-friendliness: switching off website, functions related to security 'You cannot undo this sharing even after unwanted features helps to prevent Fig. 3: A fi gure showing the structure of a the wastage of electrical, electronic, simple MASS-based system account deletion. Share with care.' and computational resources-this is General Characteristics good for economy and environment. will be spread across many PHP fi les, Previous sections provided a detailed 5. Unsolicited newsletters and Spam as many API functions, probably mixed view of MASS. Here are the general messages can be controlled. with resources related to other layers. characteristics of a social networking 6. User can 'escape' from social There will be no clear picture of a security service which has implemented MASS: networking at any time. layer's existence. But a request from a • The user can choose which features Cons client computer (i.e., a user's request to are to be enabled and not. 1. The user has to switch on features load a page, write a comment, etc.) is • A feature doesn't trigger another one manually, and in order make this always passed through a function which automatically (e.g.: a profi le page is easier, the developer has to write invokes many other security-related API not created automatically when one extra code. functions. Hence there is a solid security tries to start a blog page). 2. Client-side encryption feature asks layer existing. • User's public news feed/timeline for additional coding. Everything Turned Off by Default should not include anything without 3. Strict security and privacy makes All local layers must be disabled by default explicit approval from him/her. For censoring complicated. in a MASS-based system. That is, no example, 'Example has changed his Cons one and two are developer's issues unnecessary features are enabled without profi le photo' is shown to nobody and should be taken as duty. Solution for the consent of the user (even basic things like else until Example commands to the third one is discussed under subtitled profi le pages are considered unnecessary). do so. This may become a tiring Privacy vs Safety. A service panel can be provided which is job sometimes. So a feature can be Implementation empty by default and where the user can added which helps the user to list add the features (s)he wants. things which can appear in his/her Layout of an Example System timeline without further approval. A Figure 3 shows the structure of an example Everything Private by Default 'starred (trusted)' contact list can MASS-based system. Security and privacy When a user tries to post something, its also be created for this. feature is made global. A user has to pass visibility should be set to 'private' (or at • Each feature (e.g.: group discussion, through this layer in order to create a least 'friends') by default. There had been microblogging) is clearly new account or browse visible-for-public issues of overcrouding in many functions distinguished from one another pages. (S)he has to pass through an since people accidentally posted the invitations publicly. Accident exposure of so that the user can control them additional layer for authentication in order private data is a major privacy issue also. separately. to do microblogging or podcasting since • System-wide security and privacy both can be done only by a registered user. feature which ensures the absence of If the podcasted media should be available surveillance traps. for non-registered users (like in YouTube), • Client-side (end-to-end) encryption they have to be moved into a layer which is for private data and messages. not nested under the authentication layer. • Profi le about a user's tastes, if Developer View Point necessary, is created by asking him/ Although the theory of MASS-based her questions directly. Watching or systems gives a clear picture of properly recording the user's actions for this divided subsets, the coding need not be purpose is not allowed. this much distinguishable. Putting each • Non-addiction features. layer into its own subfolder/source fi le • The user gets the freedom to delete will not be possible always. The division is the account whenever (s)he wants. done via invoking the functions properly. Access to this feature should not That means, the idea of layers is made be hidden somewhere, but easily practical in a theoretical way rather than Fig. 4: A sample service panel accessible. physical.
CSI Communications | December 2014 | 34 www.csi-india.org Privacy vs Safety and tastes of both parties in a healthy way. websites, which sometimes resembles Strict privacy and freedom sometimes Many features are turned on Multilayer Social Networking. For example, make it diffi cult to apply community by default in most social networking blogging and podcasting services from guidelines and parental features. MASS websites. Users face diffi culties when they Google are distributed into two diff erent can have exceptions for minors making fi nd no way to turn them off . Annoying websites: Blogger.com and YouTube.com. their parents able to watch them (upon notifi cations is also a part of this issue. Google doesn't force a user to use both the basis of prior request). Only non- Unsolicited Newsletters linked. MASS supports cross-website private activities can be monitored since Many newsletter and mailing list services social networking, but it requires each the private ones might be encrypted. Anti- allows the admin to add people to the website to be compliant with the idea. social or anti-national activities can be recipients' list directly. They may get Conclusion banned on the base of manual reports. the opportunity to unsubscribe, but the Even though social media is very useful, If software daemons are running to problem is, subscription doesn't need any most social networking services are watch unethical moves, only users who work confi rmation. This happens in the fi eld threatening users' security, privacy for terrorism should be noted for manual of e-mail marketing mainly. Moreover, and health. They waste a lot of human actions. Users who do unethical things like The admin or a member can access the resource/energy which can be used spreading malware can be detected while recipients' list which is a major privacy in productive ways. So a new social passing through the security layer and they issue. In MASS, even admin cannot view networking strategy is needed and MASS can be banned without manual actions. the entire recipients' list. is introduced to fi ll that gap (at least to start such a standardisation). The (Manual action means other human beings Forbidden Layers ultimate goal of MASS is to make a social know 'this' person does 'that'.) However, Almost all websites keep restricted data nothing from the daemons should be networking service secure, private and outside the 'public_html' folder, which is the healthy. Prevention of wastage of human recorded or transmitted except reports basic idea of forbidden layers. But MASS related with terrorism. energy is also a goal of MASS. asks for more. Any user upload should be Some characteristics of MASS MASS at Present stored in a forbidden layer (intermediate can already be found in existing social Although the term MASS is being defi ned steps can use public_html directory if networking services, but of course, they are here for the fi rst time, some aspects of it have needed, given the data is erased soon), not designed in a way to be truly compliant been implemented already. For example, until the user makes it public. API source with MASS. Even though MASS makes almost all sites validate requests using code should also be kept in forbidden social media non-addictive, it will not fi lters and pass them through anti-malware layers. These actions will prove useful when discourage people from using such services. features. This is an example for global layers. things like server misconfi guration or bad Instead, a social networking service which But most of these are done considering the permissions occur. has implemented MASS will become facts like the service provider's security Cross-website Case trustworthy and encourage more people to and commercial success, where MASS is Some companies provide their social join it and start utilizing it for good purposes. introduced to protect the privacy, security networking facilities in diff erent domains/ n
Nandakumar Edamana is a first year B.Sc. Computer Science student at College of Applied Science, Vattamkulam, run by Institute of Human Resource Development, Kerala. Being a free software (free as in freedom) user, developer and activist, he has created a handful of software packages, which are available at nandakumar.co.in. Sammaty Election Engine, one of the packages developed by Nandakumar, became a popular choice among thousands of schools in Kerala. He is a frequent contributor to mainstream print media also. About the Author
CSI Communications | December 2014 | 35 Practitioner Amitava Nag Workbench Assistant Professor, Head in Dept. of IT, Academy of Technology, West Bengal
Programming.Tips() » Fun with Bitwise Operators in C Programming
The commonly used “bitwise operators” in C are: “~”, “&”, “|”, printf("\n\t %d is power of 2",n); “^”,”<<” and “>>”. These operators have enormous power. Bitwise else operators work on each bit(1 or 0) of data. Thus bitwise operators printf("\n\t %d not power of 2",n); make processing faster. The following examples (program to check return 0; whether a number is power of 2 or not) show that how bitwise } operators reduce computational cost of a program. Output: Function 1: Enter a number to check whether it is power int isPowerOfTwoWB(int n) of 2: 16 { 16 is power of 2. int flag=0,c; while(n>1) Enter a number to check whether it is power { of 2: 24 c=n%2; 24 not power of 2. if(c==1) { Another funny C program to obtain binary equivalent of a decimal flag=1; number using bitwise operator is given below: break; int main() } { n=n/2; int i,n,m,x; } printf("\t Enter A Number"); return flag; scanf("%d",&n); } printf("\ The binary equivalent of %d is ",n); If n = 1024, then while loop will be executed 10 (log2(1024)) times. for(i=15;i>=0;i--) Thus time complexity of isPowerOfTwoWB is O(log2n). On the { other hand, the following function (isPowerOfTwoB) written m=1< Amitava Nag is working as an Assistant Professor and Head in Dept. of IT, Academy of Technology, West Bengal,India and member of CSI, IEEE and ACM. He is one of the authors of the books 'Data Structures and Algorithms Using C, Numerical Methods and Programming,Basic Computation And Principles Of Computer Programming,Operating System etc. He has contributed articles to CSI Communications. About the Author CSI Communications | December 2014 | 36 www.csi-india.org Practitioner Umesh P and Silpa Bhaskaran Workbench Department of Computational Biology and Bioinformatics, University of Kerala Programming.Learn("R") » Shine with Shiny of R !!! In this session, let us jump into web app development by using R. Shiny is a wonderful RStudio package, using which you can showcase and share statistical analyses and results performed in R on the web interactively. You can work on Shiny and produce shining results, even if you don’t have prior knowledge in HTML, CSS, or JavaScript. As always, to start Shiny, you have to install and load it in the R environment. For this you can use, > install.packages(“shiny”) > library(shiny) A typical Shiny app has two components - a user-interface script and a server script. The user-interface (ui) script controls the layout and appearance of app. Server script contains the instructions for building or manipulating your data in your app. A typical user-interface script starts with a ShinyUI with Fig. 1: View of an app developed using Shiny attributes. Let us have a look into a simple user-interface script (save this as ui.R in a folder –say, myapp) Here we have used sidebarLayout to create a basic layout for your Shiny app. There are more advanced layouts according shinyUI(pageWithSidebar( to your needs. You can read more on Shiny Application Layout headerPanel(“Know Your Body Mass Index”), Guide. Also here we have used three panels - headerPanel, sidebarPanel( sidebarPanel and mainPanel for the layout. In the sidebar numericInput(‘id1’, ‘Enter Your Weight in panel we have created two numericInput which is the input Kg(id1)’, 0, min = 0, max = 10, step = 1), control for entering values and a submit button. In the main panel numericInput(‘id2’, ‘Enter Your Height in we have crated label with style h4 (the style that used to create Meter (id2)’, 0, min = 0, max = 10, step = 1), text in html). submitButton(‘Submit’) Here Shiny uses html styles to manipulate text styles. Table 1 ), (from Shiny web site) shows diff erent text styles and their usage. mainPanel( Shiny function Usage h4(‘You entered Your Height as’), p A paragraph of text verbatimTextOutput(“oid1”), h1 A fi rst level header h4(‘You entered Your Weight as’), h2 A second level header verbatimTextOutput(“oid2”), h3 A third level header h4(‘Your Body mass Index is’), h4 A fourth level header verbatimTextOutput(“oid3”) h5 A fi fth level header ) h6 A sixth level header )) a A hyper link br A line break (e.g. a blank line) Now let us see the server script (save this as server.R in the same div A division of text with a uniform style folder –myapp) span An in-line division of text with a uniform style shinyServer( pre Text ‘as is’ in a fi xed width font function(input, output) { code A formatted block of code output$oid1 <- renderPrint({input$id1}) img An image output$oid2 <- renderPrint({input$id2}) strong Bold text output$oid3 <- renderPrint({input$id1/ em Italicized text HTML Directly passes a character string as HTML code (input$id2*input$id2)}) Table 1: Text styles and their usage in Shiny } ) Also verbatimTextOutput is used to render a reactive Now run the app by typing shiny::runApp(path to the output variable as text with a variable Id – “oid1” etc. files ui.R and server.R) In the server script, renderPrint function is used to For example, shiny::runApp(‘C:/Users/DCB/Desktop/myapp’) capture the output. You can try more experiments in Shiny for which tutorial Now let’s have a closer look into the code. The fi rst line in the is available in the Shiny web page. Please have a look at user interface script shinyUI(pageWithSidebar()sets the http://shiny.rstudio.com/ tutorial and try more exciting Shiny scripts. UI for the Shiny page. We are sure that you can shine in web world by using Shiny. n CSI Communications | December 2014 | 37 Samriti Gupta*, Balvir Kumar** and P. K. Khosla*** Security Corner *Scientist C, Terminal Ballistics Research Laboratory, Chandigarh DRDO **Scientist E, Terminal Ballistics Research Laboratory, Chandigarh DRDO *** Scientist G, Terminal Ballistics Research Laboratory, Chandigarh DRDO Information Security » An Overview of Next Generation Firewalls (NGFW) Just as antivirus software has been a security threats, some non-productive meaning administrators can now easily cornerstone of PC security since the applications drain bandwidth and mine the traffic analytics to perform early days of the Internet, firewalls productivity, and compete with mission- capacity planning, troubleshoot have been the cornerstone of network critical applications for precious network problems or monitor what individual security. The previous generation of bandwidth. Importantly, enterprises employees are doing throughout the firewalls includes packet filtering, need tools to guarantee bandwidth for day. application proxies, stateful packet critical business relevant applications Most of things mentioned above inspection firewall categories. This and need application intelligence and about NGFW, make them sounds similar generation of firewalls addressed control to protect both inbound and to UTM (Unifi ed Threat Management). security in a world where malware was outbound flows of traffic, while ensuring But next generation fi rewalls are diff erent not a major issue and web pages were the velocity and security to provide a from unifi ed threat management (UTM). just documents to be read. Ports, IP productive work environment. UTM products are basically stateful addresses, and protocols were the key Next-generation fi rewalls (NGFW) inspection fi rewalls with some additional factors to be managed. Network security represent the next major step in the security functionality. These products used to be relatively simple; everything development of fi rewall technology in often consolidate fi rewall, intrusion was more or less black and white (either this direction. A next-gen fi rewall is prevention, content fi ltering, antivirus clearly bad or clearly good). Problems designed to combine the functionality and other security functionality into a with this approach today include the of a fi rewall and an IPS, while adding single box. While this approach is not fact that applications have become detailed application awareness into the often appropriate for a large enterprise, increasingly “gray”, now classifying mix. NGFWs bring additional context to a UTM device is eff ective product for types of applications as good or bad are the fi rewall’s decision-making process smaller or midsize enterprises seeking not a straightforward exercise. by providing it with the capability of to limit security expenditures. Signifi cant As the Internet evolved, the ability understanding the details of the Web functional requirements for an eff ective to deliver dynamic content from the application traffi c passing through it, next-generation fi rewall over UTM include server and client browsers introduced a taking action to block traffi c that might the ability to: wealth of applications we now call Web exploit Web application vulnerabilities. • Identify applications regardless 2.0. Port-based fi rewalls are helpless in Next-generation fi rewalls combine of port, protocol, evasive providing security as today, applications the capabilities of traditional fi rewalls techniques, or SSL encryption from reputed to nefarious all run over - including packet fi ltering, network before doing anything else. TCP port 80 as well as encrypted SSL address translation (NAT), URL blocking • Provide visibility of and granular, (TCP port 443). Additionally, for want of and virtual private networks (VPNs) - with policy-based control over easy availability and cost savings to the Quality of Service (QoS) functionality applications, including individual business, many client-server applications and features not traditionally found in functions. like Salesforce.com and Google’s Offi ce fi rewall products. These include intrusion • Accurately identify users and Suite are moving to the web to become prevention, SSL and SSH inspection, subsequently use identity web-based services. Such critical deep-packet inspection and reputation- information as an attribute for business applications have today become based malware detection as well as policy control. indistinguishable from the less important application awareness. The application- • Provide real-time protection applications in a business network that specifi c capabilities are meant to thwart against a wide array of threats, also utilize HTTP for the purpose of the growing number of application including those operating at the network communications. attacks taking place on layers 4-7 of the application layer. These scenarios lead to a situation OSI network stack. • Integrate, not just combine, where Enterprises, need a deeper A next-generation firewall inspects traditional firewall and awareness of and control over individual the payload of packets and matches network intrusion prevention applications along with deeper signatures for nefarious activities capabilities. inspection capabilities by the firewall such as known vulnerabilities, exploit • Support multi-gigabit, in-line that allow administrators to create very attacks, viruses and malware all on the deployments with negligible granular allow/deny rules for controlling fly. There are some indirect benefits of performance degradation. use of websites and applications in this also; since the contents of packets The last point mentioned is quite the network. This is also necessary are inspected, exporting all sorts of significant, as this distinguish NGFW because in addition to introducing statistical information is also possible, from the cadre of normal firewalls. For CSI Communications | December 2014 | 38 www.csi-india.org traditional security products, especially reduces hard capital costs, as well as secured using NGFWs. But the capabilities those with bolted-on capabilities, ongoing “hard” operational expenses, that they have, there future is bright in each high-level security function is such as support, maintenance, and cyber market and it is hoped that by the performed independently. This multi- software subscriptions, power and HVAC, end of 2014, their usage will rise to 35% pass approach requires low-level packet and “soft” operational expenses, such as of the installed base, with 60% of new processing routines to be repeated training and management. purchases being NGFWs. numerous times. System resources are IT and business alignment: Enable IT to used inefficiently and significant latency confidently say “yes” to the applications References is introduced. In contrast, a NGFW that needed to best support the business — [1] Intrusion Prevention Systems (IPS): uses single-pass architecture eliminates by giving them the ability to identify and Next Generation Firewalls, A Spire repetitive handling of packets, reducing granularly control applications while Research Report, By Pete Lindstrom, the burden placed on hardware and protecting against a broad array of Research Director. minimizing latency. threats. [2] Improving Network Security: Next Next-generation fi rewalls produce NGFW are quite effective than Generation Firewalls and Advanced numerous benefi ts over traditional previous generation firewalls, but there Packet Inspection Devices, By network security infrastructures and some things to keep in mind to evaluate Steven Thomason, Global Journal of solutions. These include and embrace a next-generation firewall. Visibility and control: The enhanced As a baseline, there must have a thorough Computer Science and Technology visibility and control provided by NGFWs understanding of organization’s needs Network, Web & Security. enable enterprises to focus on business and should have performed extensive [3] An Overview of Next-generation relevant elements such as applications, testing before deciding to implement Firewalls, http://www.techrepublic. users, and content for policy controls, NGFW. There are few things that com/blog/it-security/an-overview- instead of having to rely on nebulous should be looked into before deploying of-next-generation-fi rewalls/ and misleading attributes like ports NGFW, like ease of the use managing [4] Next Generation Firewalls for and protocols, and to better and more interface, vendor specific application Dummies, By Lawrence C. Miller, thoroughly manage risks and achieve identification support, performance CISSP, palo alto Networks. compliance, while providing threat characteristics in real environment as [5] Next Generation Firewalls Security prevention for allowed applications. sometimes the performance of NFGW Without compromising Performance, Safe enablement: Achieve changes drastically by the amount of http://www.techrepublic.com/blog/ comprehensive coverage — by, providing traffic or switching some application it-security/next-generation-fi rewalls- a consistent set of protection and inspection functionality. enablement capabilities for all users, Due to the various elements that security-without-compromising- regardless of their location. need to be considered, and also because performance/ Simplifi cation: Reduce complexity of the of the relative newness of the technology, [6] Next Generation Firewall, http:// network security and its administration NGFWs are not yet widely adopted by www.niiconsulting.com/solutions/ — by obviating the need for numerous organizations. According to a survey, less next-generation-fi rewalls.html stand-alone products. This consolidation than 5% of Internet connections today are n Samriti Gupta is working as Scientist C in Terminal Ballistics Research Laboratory Chandigarh, DRDO. She has done M.E. in Computer Science & Engineering. She has more than 6 years of scientifi c research experience. She is Associate Member of Aeronautical Society of India and recipient of team award of DRDO, Agni Excellence Award for Strategic Contribution. Her research interests include Cyber Security, Information Security, Embedded Systems etc. Balvir Kumar is working as Scientist E in Terminal Ballistics Research Laboratory Chandigarh, DRDO. He is M.Tech from IIT, Madras in Computer Science & Engineering. He has more than 14 years of research experience. He has been the recipient of DRDO Young Scientist award of the year 2012. He has been working in the area of Embedded Systems, Computer Architecture, Networking etc. P. K. Khosla is working as Scientist G in Terminal Ballistics Research Laboratory Chandigarh, DRDO. He has completed M.Tech with university gold medal. He has more than 27 years of research experience in various domains including Telemetry Systems, Embedded Systems, Networking and supersonic track testing. He has been a recipient of number of technical awards and has the credit of publication of research papers in various journals and international conferences. About the Authors CSI Communications | December 2014 | 39 Dr. Vishnu Kanhere Security Corner Convener SIG – Humane Computing (Former Chairman of CSI Mumbai Chapter) Case Studies in IT Governance, IT Risk and Information Security » Algorithmic Computing – Problem or Solution? Algorithm, simply put is the technique used to get a job done. It is a method to solve a problem that consists of exactly defi ned instructions. Some tasks are one-off problems. One thinks through the solution, applies it and then just moves on. Most jobs, tasks and problems are apparently similar and are faced and dealt with again and again. It is for these that you need a solution, an approach that needs to work every time without having to think through the tasks and solutions. This makes life easier for all concerned - be it a practicing professional, an entrepreneur, an employee or a common person going about his day to day life. Use of algorithmic computing has enabled today’s advanced decision support systems, core banking solutions, auto-pilot mechanism in aircrafts, case management software for doctors and lawyers, and the smart washing machine and smart television set both of which remember your personal choices and preferences – e.g. decide the duration of wash and spin cycle based on the clothes loaded etc. Modern digital cameras use algorithmic programming to ensure that pictures are sharp and focused and use optimum eff ects. There are problems too. Systems often act dumb and lack common sense, which an ordinary human being of average intelligence also displays and thus the individual is able to deal with situations better than say a computer program. One still comes across massive utility bills charged on inoperative / low use accounts which are attributable to program malfunction. No doubt, algorithmic programming is here to stay and makes programs and computers better, faster, more effi cient, more eff ective, and most important is that it takes away the load of repetitive tasks and problems from humans to systems. In fact the prowess of computing systems in terms of speed, reasoning and responses has been such that computers / systems are often regarded as a panacea for business and industry problems and tasks. CRM software, decision support systems, Algo-trading software for dealers and brokers on the stock market are a few examples of systems designed using algorithmic programming. To elaborate, an automated (Robotic) trading system also referred to as Algo trading, automatic trading, algorithmic trading or automatic trade executor, allows traders to enter and exit the trade without human intervention based on simple or complex conditions. Most of the Robotic Trading systems require charting software with real time data, trade executor plug-in and brokers trading terminal. The fi rst thing in Robotic Trading is to create a trading strategy. The strategy can be moving average crossover, pivot level break, opening range breakout (ORB) or oscillator related strategies. Based on this high volume, multiple automated trades are placed. Based on the exit strategy, it will book profi t or exit when stop loss hits. The system utilizes highly advanced mathematical formula to generate the trading decisions in equity and derivative markets. Algo trading or robotic trading is generally used by institutional traders who have heavy volume trading to automate their work, but now even small traders have started using robotic trading software. What can go wrong or what issues could arise using such systems more frequently and on a larger scale? Is it really safe to depend on systems for important business, welfare / social / health and personal decisions? These are signifi cant issues especially in the context of Governance, and Security. Businesses and society need to consider these issues before progressively adopting and relying on these systems, but we have gone so far ahead that in reality there is no turning back; it is a point of no return. It is necessary to look into these aspects more closely. Given this background the current Case in Information Systems is being presented. The facts of the case are based on information available in media reports online and real life incidents. Although every case may cover multiple aspects it will have a predominant focus on some aspect which it aims to highlight. A case study cannot and does not have one right answer. In fact answer given with enough understanding and application of mind can seldom be wrong. The case gives a situation, often a problem and seeks responses from the reader. The approach is to study the case, develop the situation, fi ll in the facts and suggest a solution. Depending on the approach and perspective the solutions will diff er but they all lead to a likely feasible solution. Ideally a case study solution is left to the imagination of the reader, as the possibilities are immense. Readers’ inputs and solutions on the case are invited and may be shared. A possible solution from the author’s personal viewpoint is also presented. A Case Study of Vaayda Bazaar Ramesh, the CEO of Vaayda Bazaar has • Earn profi t from Arbitrage which can substantially fund the acquisition decided it’s high time to introduce Algo opportunities between multiple of the system. It is amidst these confl icting trading in the company. After his recent exchanges opinions in the top management team that US trip where he has seen these systems • Curtail broker front-running and a go ahead is given and a time frame of 6 work at lightning speed placing spliced impact costs months is given to implement and roll out orders and reaping in arbitrage profi ts • Allow purchases to be spread over the system in the company. he is convinced that in these days of thin number of smaller trades preventing Prabhudas is concerned and he has margins and online trading platforms prices from being driven up pointed out two signifi cant reports adopting Algo trading is the way forward • Enhance liquidity and also The fi rst is about the 2010 US fl ash for the company. • Enable large orders otherwise crash that happened in May 2010: He calls a meeting of his management blocked by the exchanges. A computer-driven sale worth $4.1 team and shares his viewpoint. He Prabhudas, who heads market billion by a single trader helped trigger the emphasizes that adopting Algo trading operations is an old timer. He and May fl ash crash, setting off liquidity shocks will enable Vaayda Bazaar to – the chief accountant Kantilal are not that ricocheted between U.S. futures and • Open up and capitalize on great enthused. The CIO Kunal is bullish and is stock markets, regulators concluded in a trading opportunities with less all set to implement the system. In fact he report. The report by the U.S. Securities human error anticipates a fair reduction in staffi ng costs and Exchange Commission and the CSI Communications | December 2014 | 40 www.csi-india.org Commodity Futures Trading Commission Ramesh who had already made up of a situation and that would be did not name the trader. Reuters, citing his mind, pointed out that at this stage extremely diffi cult to predict. internal documents prepared by exchange the debate was only academic and inputs The Consequences operator CME Group Inc, in May identifi ed provided by Prabhudas were exceptions to The company has to be alive to these the trader as money manager Waddell the rule and could be used benefi cially as issues when implementing and using & Reed Financial Inc. The long-awaited inputs to improve the system. automated algorithmic trading. The report focused on the relationship To reassure the old guard, Ramesh company its management and employees between two hugely popular securities decides to call in Sulabha an experienced cannot individually and collectively ignore - E-Mini Standard & Poor's 500 futures analyst who had implemented similar these issues. Doing so can not only result and S&P 500 "SPDR" exchange-traded systems to help them move to Algo in losses but even wipe out the existence of funds - and detailed how high-frequency trading. Sulabha has been called and she the company, in a single day. algorithmic trading can sap liquidity and has suggested a way forward. The Strategy rock the marketplace. "The interaction Solution The right strategy for the company will be between automated execution programs The Situation to deal with each of these issues: and algorithmic trading strategies can High frequency trading using automated 1) Bugs in the software – governance quickly erode liquidity and result in software programs is dominating world and control on software acquisition, disorderly markets," the report said. The stock markets today. It accounts for up to development, deployment and use "fl ash crash" sent the Dow Jones industrial 50% of the trade volume and has become needs to be put in place and rigorously average plunging some 700 points in acceptable and is used across all major monitored. UAT, and regular periodic minutes on May 6, exposing fl aws in the exchanges. It enables fi rms both large testing, mock drills in simulated electronic marketplace dominated by and small to use automated strategies conditions and audit post changes high-speed trading. The Dow was down to churn through large volumes of orders with a proper change management nearly 1,000 points at its lowest point on in fractions of seconds. Some fi rms can procedure needs to be put in place. that day.” trade in microseconds. After introduction 2) Human errors – Human errors need The second one was the NSE October of online trading 10 to 15 years back, this is to be restricted at the programming 2012 crash that wiped nearly $60bn off the next big change across markets. level through robust software Indian stock index. It has increased volumes, lowered development and in data entry and India's main share index plunged arbitrage, made markets more liquid and use through appropriate controls to 16% within minutes on Friday in a so- has taken some of the emotion out of the prevent these errors. Further, putting called "fl ash crash" – the latest in a series markets making it more rational. Clearly, in place risk management controls of market glitches that have shaken it provides an advantage to the fi rms that in the form of inbuilt checks like confi dence in global fi nancial markets. use it, and is here to stay. quantity and value limits, price range India's National Stock Exchange (NSE) Given its size and scale of operations checks, client level, broker level and was forced to halt trading for 15 minutes, Vaayda Bazaar has to introduce Algo overall exposure and credit limits, will after a brokerage placed 59 wrong orders, trading and automate its systems. The key minimize the impact of human errors. triggering a sell-off that wiped nearly challenge is identifying what can go wrong 3) Unauthorized changes / alterations $60bn off the value of the country's which can then be dealt with eff ectively. and Rogue elements – Information biggest companies. The orders were • The fi rst issue is of bugs in the security governance framework cancelled and stocks recovered, with software – programming errors, using proper policies, best practices, India's Nifty index closing down 0.8% on errors in logic, in the strategies built procedures, controls and monitoring the day. NSE blamed Friday's crash on into the system. needs to be implemented and human error but experts said that was • The second issue is of human errors maintained to secure the system. highly unlikely. "There is no human being in programming as well as in feeding 4) Black Swan – Dealing with the black in the world that can take down the stock in, selection and use of the strategies swan is the biggest challenge as market by 16%. This is typical spin." They in given situations. algorithmic programs cannot deal said the error was exacerbated by high- • The third issue which is of greater with them, simply because they are frequency trading, which involves using concern is of unauthorized changes / not designed for them. In fact just like software to post orders for microseconds alterations to the program which can people they are blind to uncertainty at a time to exploit tiny diff erences in cause huge losses to the fi rm. These and unaware of the massive role of share prices. It is now widely accepted could be motivated by fraud or malice the rare event in historical aff airs. that high-frequency trading fuelled the and may be internal or external. They only deal with and are based on fl ash crash of May 2010, which saw the • The fourth is the possibility of events that have occurred and have Dow Jones industrial average plunge by rogue traders or elements using the been encountered in the past, that 998 points in 20 minutes, raising fears of systems setting off a market crash. too in the recent past. a worldwide stock market collapse. • Finally, the most disturbing of all Black swan events are typically random Kunal points out that computers do is the occurrence of what Nassim and unexpected. For example, the previously exactly as they have been programmed to Nicholas Taleb, a fi nance professor successful hedge fund Long Term Capital do, and there will be large errors only if they and former Wall Street trader had Management (LTCM) was driven into aren't monitored properly. According to him described as the “Black Swan eff ect”. the ground as a result of the ripple eff ect all these were human errors and blaming the An event or occurrence that deviates caused by the Russian government's debt Algo systems was just witch-hunting. beyond what is normally expected CSI Communications | December 2014 | 41 default. The Russian government's default who have experience and understanding leaving no choice to the risk manager represents a black swan event because of the market. This may not prevent losses who has to be always alert and on the none of LTCM's computer models could but will at least ensure that the program lookout. have predicted this event and its subsequent which is blind to the rare event can be The company will do well to put in eff ects. stopped in time preventing a complete place these measures initially and keep The best way to deal with such wipe off . This will enable the company to identifying newer threats and issues as eventualities would be to put in place survive another day. they emerge and follow best practices to risk management systems which are In modern times with transaction make itself Algo trading ready. monitored and provide for human speed being in micro seconds the An eff ective solution is generally intervention at a very senior level by people fuse wire of decisions is really short – expected to proceed on these lines. n Dr. Vishnu Kanhere Dr. Vishnu Kanhere is an expert in taxation, fraud examination, information systems security and system audit and has done his PhD in Software Valuation. He is a practicing Chartered Accountant, a qualifi ed Cost Accountant and a Certifi ed Fraud Examiner. He has over 30 years of experience in consulting, assurance and taxation for listed companies, leading players from industry and authorities, multinational and private organizations. A renowned faculty at several management institutes, government academies and corporate training programs, he has been a key speaker at national and international conferences and seminars on a wide range of topics and has several books and publications to his credit. He has also contributed to the National Standards Development on Software Systems as a member of the Sectional Committee LITD17 on Information Security and Biometrics of the Bureau of Indian Standards, GOI. He is former Chairman of CSI, Mumbai Chapter and has been a member of Balanced Score Card focus group and CGEIT- QAT of ISACA, USA. He is About the Author currently Convener of SIG on Humane Computing of CSI and Topic Leader – Cyber Crime of ISACA(USA). He can be contacted at email id [email protected] CSI Communications | December 2014 | 42 www.csi-india.org Dr. Debasish Jana Brain Teaser Editor, CSI Communications Crossword » Test your Knowledge on Algorithmic Computing Solution to the crossword with name of fi rst all correct solution provider(s) will appear in the next issue. Send your answers to CSI Communications at email address [email protected] with subject: Crossword Solution - CSIC December 2014 CLUES ACROSS 2. A technique that converts a data string into a numeric string output of fi xed length (4, 9) 3. Type of a problem that can be solved in polynomial time (9) 6. Tree traversal technique that processes all subtrees recursively and fi nally processes the root (9) 8. A nonempty, proper subset of vertices of a graph (3) 9. A path in a graph where edges may be repeated (4) 11. A collection of items (4) 13. The maximum of the distances between all possible pairs of vertices of a graph (8) 14. An edge of a connected graph whose removal would make the graph unconnected (6) 15. A technique used in discrete and combinatorial optimization problems (6, 3, 5) 16. An array data structure that stores bits compactly (3, 6) 18. The computer scientist who formulated the dining philosophers problem (6, 8) 23. Transition systems that accept or reject their inputs, depending on the state (6, 8) 24. The number of edges connected to a vertex of a graph (6) 26. The measure of execution time of an algorithm in terms of input size (4, 10) 27. A type of pseudo-random number generator technique (6, 12) 29. Tree traversal technique that processes the root fi rst and then all subtrees recursively (8) 30. Technique that fi nds the position of an element within a sorted array (6, 6) 31. A lookup table that is equivalent to any logic circuit or loopless transition system (5, 5) 32. A rearrangement of elements of a sequence, where none is lost, added, or changed (11) 33. A string matching algorithm (5, 5) DOWN 1. A queue where request with highest priority is processed fi rst (8, 5) 4. A strategy for searching in a graph (7, 5) 5. The boolean and function (11) 7. An arrangement of characters and symbols expressing a pattern (7, 10) 8. Choosing a subset of m elements out of n elements, where m ≤ n (11) 10. Type of technique that always takes the best immediate solution while fi nding an answer (6) 12. A sorting technique (6, 4) 13. The study of mathematical structures that are fundamentally discrete rather than continuous (8, 11) 17. A greedy technique in graph theory that fi nds a minimum spanning tree for a connected weighted graph is known as ___'s algorthm (7) 19. A node in a tree without any children (4) 20. The beginning characters of a string (6) Did you ever visit Alan Turing Memorial inEngland? 21. Division into many distinct classifi cations (11) 22. A theoretical computing machine invented by Alan Turing to serve as an ideal model for Alan Mathison Turing (1912-1954), universally recognized as mathematical computation (6, 7) the father of modern Computer Science, formalized the 23. A set of strings of symbols that may be constrained by rules those are specifi c to it (6, 8) concepts of algorithm and computation with the Turing 24. The boolean or function (11) machine, which has been considered a model of a general 25. A closed, bounded N-dimensional fi gure whose faces are hyperplanes (8) purpose computer. A statue of this icon of computing 28. A set of statements that defi nes a sequence of operations on each input, and one that resides in Sackville Park in Manchester, England and the eventually halts on each input (9) Turing memorial plaque in the park is shown here that reads a Bertrand Russell quote that "Mathematics, rightly viewed, possesses not only truth, but supreme beauty — a beauty cold and austere, like that of sculpture." Solution to November 2014 crossword (More details can be found in http://en.wikipedia.org/wiki/Alan_Turing and http://en.wikipedia.org/wiki/Alan_Turing_ Memorial) We are overwhelmed by the responses and solutions received from our enthusiastic readers Congratulations! NEAR ALL correct answers to November 2014 month’s crossword received from the following readers: Er. Aruna Devi (Surabhi Softwares, Mysore), Dr. Suresh Kumar (CSE, Faculty of engineering and Technology, Manav Rachna International University, Faridabad) and S Christina Mary and S Jeyalakshmi (MSPVL Polytechnic College, Pavoorchatram, Tamil Nadu) CSI Communications | December 2014 | 43 H R Mohan Happenings@ICT ICT Consultant, Former AVP (Systems), The Hindu & President, CSI Email: hrmohan.csi@gmail .com ICT News Briefs in November 2014 The following are the ICT news and servers are located outside India. There Company News: Tie-ups, Joint Ventures, headlines of interest in November 2014. are around 40 million such websites New Initiatives They have been compiled from various around the world – IAMAI. • e-toll collection (ETC) in 350 national news & Internet sources including the Govt, Policy, Telecom, Compliance highways by December. ETC could save Rs. 34,000 crore - Nitin Gadkari. • IT Ministry to expand the incubator dailies - The Hindu, Business Line, and • Google launches the Indian Language infrastructure for start-up fi rms off ered Economic Times. Internet Alliance (ILIA) to promote Indic by the STPI from four lakh sq ft at present language content online. Voices & Views to 10 lakh sq ft. It plans to start a ‘trade • Silicon Valley start-up develops intelligent • The IT-BPM industry contributed 8.1% of net portal’ to provide B2B services to the sound system for iPhones. country’s GDP last year and garnered $86 IT industry. • The Silicon Valley chapter of TiE to launch billion in terms of revenue exports. It has • India, Australia likely to ink social security ‘Billion Dollar Babies’ to identify promising given direct employment to 3.1 million and pact which will benefi tIT professionals tech start-ups that have the potential to 10 million as indirect employment. going for short visits. reach a billion dollars in valuation. • “Indians have it in their DNA to build the • Over 1.25 crore pensioners in the country • India is home to some 3,100 start- next WhatsApp for the world.” - Neeraj will no longer be required to submit ups. Nasscom to take 150 of the most Arora, Vice-President of WhatsApp. physical proof of being alive every promising tech start-ups to Silicon Valley • About 53 per cent of children in India have November, thanks to an Aadhaar-based in 2015 to showcase their work and talent. been cyber bullied – Study by Telenor & digital certifi cation system soon. • Cognizant pips Infy, Wipro in revenue BCG. • Next round of auction could force growth in Sep 2014 quarter. • There are 30 IT software product operators out of business in some circles - • WhatsApp, the messenger service fi rm companies in India with a combined Vodafone India head. acquired by Facebook, is not to enforce its valuation of a whopping $6.2 billion or • A special Defence Interest Zone to be $1 a year fee in India. Rs. 37,500 crore. created near international border for war • Bharti Airtel launches a free WiFi service • Indian Gaming market is worth close to time and counter- terrorism operations. in two West Bengal Government-run AC $890 million and casual gaming revenue • Masking phone numbers will need buses in Kolkata. is set to quadruple from $65 million to security clearance. • India, Russia look to establish smart cities. $244 million by 2015. • BSNL-Inmarsat satellite telecom services • Singapore keen to develop AP capital • Healthcare providers will spend $1.1 billion set to take off at a cost of $8 million with including development of a new smart on IT in 2014 – Gartner. Ghaziabad as the location for the satellite satellite city. • Jack Ma-promoted Alibaba, the world’s gateway. • Ramco Systems has developed ERP largest e-commerce company by value • DoT has identifi ed 45 cities and 705 software suitable for use on wearable and volume, mopped up a whopping tourist locations for Wi-Fi services in devices such as Google Glass and smart $8 billion (Rs. 48,000 crore). public areas. watches. • India’s online retailing is expected to reach • To help IT fi rms, India will seek more visa • ICAR - Indian Council for Agricultural Rs. 50,000 crore by 2020. concessions from US. Research, brings in IT tools to forecast • Out of the 48 million SMBs in India, • A super-regulator for broadcasting, IT and commodity prices. only 5-6% businesses have a website telecom through a new Communication • Bangalore to host maiden CeBIT in which whereas. 40% of the SMB owners have a Bill will repeal four existing Acts, including 50 start-ups to get special display. smartphone. the Indian Telegraph Act 1885 and the • American Tower keen to develop ‘villages • Bangalore will become the world’s largest Telecom Regulatory Authority of India of the future’. IT cluster by 2020. Bangalore contributed Act 1997. • BIAL becomes the fi rst e-freight $45 billion, or 38% of India’s total IT compliant airport in India. It eliminates exports, including domestic consumption, IT Manpower, Staffi ng & Top Moves the generation of up to 30 documents an last fi scal. - IT Secretary, Karnataka • 8 years, 800 million handsets later, average airfreight shipment generates. Government. Nokia shutters Chennai plant. Last 800 • NIIT, eBay join hands for e-commerce • The IT industry must change to the employees took VRS. certifi cation programme. entrepreneurial mindset of ‘Believing • EMC has tied up with 27 engineering • IBM extends its Global Entrepreneur is Seeing’ where they believe in an colleges in Pune to train students in the Program to Indian start-ups. opportunity and see it, after which fi eld of information storage management. • SanDisk in talks with Indian phone-makers everybody else sees it – CEO, Infosys • Infy probing harassment charges by a to enhance memory of devices. • With 3,100 technology start-ups, India is group of female employees against top • Intel, unveils its “Sensing Platform”, with home to the third-largest start-up base in executive. an eye on the $7 trillion Indian healthcare the world - NASSCOM. • Karnataka alone created 73,000 new jobs market. • 108 billion work emails are sent daily, in IT, last fi scal year and employs 10 lakh • Zomato, the popular food and discovery requiring one to check inbox an average of people directly and 30 lakh indirectly. This service to launch a mobile payment app. 36 times an hour. Only 14% of the emails number is expected to double to 80 lakh • IBM launches email service – IBM Verse. are of critical importance – IT Industry direct and indirect jobs by 2020. • Microsoft, partners to launch e-learning Analysts. • Gujarat Govt aims at 1 million IT jobs by platform soon. • The market for the Internet of things is 2020. • Microsoft Devices bets on services, apps projected to hit $7.1 trillion by 2020 - IDC. • Syntel initiates a corporate recruitment to drive smartphone sales. • The online shopper base which was at drive to reach out to all eligible candidates • Cisco tests ‘Make in India’ waters with 8 million in 2012 grew to 35 million by from more than 2,800 AICTE approved $5 billion proposal. 2014. It is expected to grow to 100 million engineering colleges. • Mobile operators off er credit facilities by 2016. Over 50 million new buyers to • Wipro selects 14 students from Karnataka for pre-paid users to help them make come from Tier 1 and Tier 2 cities – Google. govt colleges under ‘Earn-cum-Learn’ important calls even after their talk-time • Total ban on porn websites not possible as plan. is exhausted fully. n CSI Communications | December 2014 | 44 www.csi-india.org CSI Reports From CSI SIG and Divisions » Please check detailed news at: http://www.csi-india.org/web/guest/csic-reports SPEAKER(S) TOPIC AND GIST SIG-FM (Special Interest Group on Formal Methods) & CSI Bangalore Chapter Dr. Kalyan Krishnamani, Prof. Deepak D’Souza, Prof. Aditya 15-17 October 2014: National Workshop/ Conference on “Formal Methods Kanade, Dr. Sriram Rajamani, Prof. Madhavan Mukund, (NCFM-2014)” Prof. Shyamsundar, Prof. Meenakshi D’Souza, Dr. Yogananda, R Venky, N Ranjana, Dr. Prahlad Sampath, Dr. Manoj Dixit, Event was fi rst of its kind. Advances in technologies, safety and security are key issues Dr. Sudarasan, Dr. Raoul Jetley, Shailashree Patil, BS Reddy and and there is need to use eff ective and reliable approaches to design, develop and Ravi Prakash qualify complex, high assurance system software within time-schedule and budget. Formal methods are proving eff ective in meeting these criteria. FMSE (Formal Methods Software Engineering) strives to promote research and development for improvement of formal methods and tools for industrial applications. Workshop had various sessions such as - Deductive verifi cation of C programs, Refi nement-Based Verifi cation, Proofs of functional correctness of data-structure implementations with illustrations, Race detection for Android applications, Probabilistic Programming, Statistical Model Checking by Prof. Mukund, Formal Methods and Web Security, Architectural semantics of AADL using Event-B, Certifying Safety Critical Control Systems – An industry perspective, Formal Verifi cation at Nvidia, Should we use formal methods in Flight Controls?, Challenges in Static analysis and testing in Industry, Validation Methods - AGNI Perspective, Formal methods at MathWorks, Formal Methods for Assurance of Safety Critical Systems: A case study, Industry perspective on use of formal methods & program analysis for embedded systems, Case Study by MCSRDC-HAL, Development of Large scale airborne embedded Inaugural session application for a fi ghter aircraft, Unleash the unknown: A true black box approach. Hi-Tech Institute of Technology, Bhubaneswar, Odisha in association with Computer Society of India (CSI) Dr. Fakir Charan Parida, Prof (Dr.) Anirman Basu, Prof (Dr.) AK 8 November 2014: Seminar on the topic “Art & Science of Love and its Nayak and Sanjay Mohapatra Management in Human Life” On this occasion, a talk was delivered by Dr. Fakir Charan Parida on the above topic. Dr. Parida is a notable person in Nuclear Science domain. Presently, he is serving as Scientifi c Offi cer in Radiological Safety Division at Indira Gandhi Center for Atomic Research, Kalpakkam (Tamil Nadu). Dr. Parida, in his talk, highlighted diff erent types and aspects of love. The nice design & contents of his talk were unique & enlightening in nature. Honouring the guest CSI Communications | December 2014 | 45 Kind Attention: Prospective Contributors of CSI Communications Please note that Cover Themes for forthcoming issues are planned as follows: • January 2015 – IT Infrastructure • February 2015 – Quantum Computing • March 2015 – Machine Translation Articles may be submitted in the categories such as: Cover Story, Research Front, Technical Trends and Article. Please send your contributions before 20th of a month prior to the issue month for which you are contributing. The articles may be long (2500-3000 words maximum) or short (1000- 1500 words) and authored in as original text. Plagiarism is strictly prohibited. Please note that CSI Communications is a magazine for membership at large and not a research journal for publishing full-fl edged research papers. Therefore, we expect articles written at the level of general audience of varied member categories. Equations and mathematical expressions within articles are not recommended and, if absolutely necessary, should be minimum. Include a brief biography of four to six lines for each author with high resolution author picture. Please send your articles in MS-Word and/or PDF format to the CSI Communications Editorial Board via email id [email protected]. (Issued on behalf of Editorial Board of CSI Communications) CSI Communications | December 2014 | 46 www.csi-india.org CSI Elections 2015-2016/2017 Following is the fi nal election slate by the Nominations Committee (2014-2015) for the various offi ces of the Computer Society of India for 2015-2016/2017 For the term 2015 - 2016 (April 1, 2015 – March 31, 2016) National Nomination Committee Vice President cum President Elect • Prof.(Dr.) Anil K Saini • Dr. Anirban Basu • Mr. C.G. Sahasrabudhe • Mr. Rajeev Kumar Singh • Dr. D D Sarma • Mr. Satish Babu • Prof. ( Dr.) Kamlesh Kumar Saini • Mr. Suresh Tiwari • Mr. S. Ramanathan • Mr. Tarun Kumar Dey • Prof. (Dr.) U.K. Singh For the Term 2015-2017 (April 1, 2015 – March 31, 2017) Hon. Treasurer Divisional Chair Person Div. I (Systems) • Prof. R K Vyas • Mr. Apoorva Agha • Mr. Soman S. P • Mr. B S Bindhu Madhava Regional Vice President (Region I) • Prof. M N Hoda • Mr. Brijendra Singh Divisional Chair Person Div. III (Applications) • Mr. Shiv Kumar • Mr. H R Raghavendra Rao Regional Vice President (Region III) • Ms. Mini Ulanat • Prof. R P Soni • Mr. Ravikiran Mankikar • Dr. Vipin Tyagi Divisional Chair Person Div. V (Education & Research) Regional Vice President (Region V) • Prof. Dipti Prasad Mukherjee • Mr. I L Narasimha Rao • Mr. Iqbal Ahmed • Mr. Raju L Kanchibhotla • Dr. P. Sakthivel • Mr. T Sabapathy • Dr. R. Rajkumar Regional Vice President (Region VII) • Dr. Rabi Narayan Satpathy • Dr. Suresh Chandra Satapathy • Mr. K.Govinda • Dr. M.A.Maluk Mohamed For the Term 2015-2017 Kolkata Chapter – Treasurer • Dr. Ajanta Das • Dr. Ambar Dutta • Mr. Sumanta Bhattacharya • Mr. Sourav Chakraborty • Dr. Abhik Mukherjee • Dr. Sanjoy Kumar Saha For the Term 2015-2016 • Mr. Devapriya Chatterjee • Dr. Tanushyam Chattopadhyay Kolkata Chapter- Management Committee • Prof (Dr.) Amitava Dass • Ms. Madhumita Sengupta • Mr. Tamal Deb • Prof. Parmartha Dutta • Mr. Radhikamohan Sanbui • Mr. Bidyut Chakraborty • Mr. Sandip Kumar Ghosh • Mr. Avijit Dharchoudhuri • Mr. Mrinmoy Chatterjee • Ms. Sharmila Ghosh • Prof. Subho Chaudhuri For the Term 2015-2017 Bangalore Chapter – Treasurer • Mr. Yogananda Jeppu • Mr. Satish B.G. • Mr. Dattatreya S Vellal • Mr. R. K. Senthil Kumar • Mr. Mohan Ramanathan • Mr. Anbunathan R For the Term 2015-2016 • Dr. Manju Nanda Bangalore Chapter – Management Committee • Dr. Prahlad Rao • Dr. CKB Nair • Dr. Arindam Sen • Dr. K. Satyanarayan Reddy • Mr. Himanshu Gupta • Mr. Ravi K. S • Mr. Dinakaran Pillai Appeal to Members You may be aware that the election for the various positions of the Executive Committee and three members of Nomination Committee for the term 2015-16/2017 as the case may be, will be through electronic ballot. • To exercise your ballot you are requested to login to http://www.directvote.net/csi using your CSI Membership Number and password. Please view your Region and Balloting options for the elective offi ces for the term 2015-2016/2017. • In case you have not received your password, please write to [email protected] on or before January 10, 2015 and we will email your login password. • Please exercise your vote for the posts of Vice President cum President elect, Nomination Committee (3 members), Hon. Treasurer, RVP-I, RVP-III, RVP-V, RVP-VII and Div- I, Div-III, Div-V. Members belonging to Kolkata and Bangalore chapters will be allowed to vote for positions in their respective chapters only. In case of RVP’S members belonging to a particular region would only vote for that region’s RVP. • The ballot also includes some amendments to the constitution (proposed by ExecCom) to be approved. You are requested to exercise your voting for these also. • The balloting for this election will start on 16th Dec, 2014 at 18.00 hrs (IST) and will end at 18.00 hrs on 16th Jan, 2015. You are requested to exercise your voting right electronically before the closure of the election site. If you have any queries on internet balloting, please email to [email protected] for clarifi cations. Code of Conduct: - Information on CD may be used for personal contacts and ethical canvassing about one’s own candidature only. The information should not be used for disturbing the privacy and disclosing the details to others or to use any derogatory/unethical remarks about the other contesting candidates. You can use SMS/ Email or Telephone for canvassing but avoid repeated messages / calls to respect privacy of the members. Thanking you and with season’ greetings Nominations Committee 2014-2015 Mr. Sanjeev Kumar Prof. P Kalyanaraman Mr. Subimal Kundu (Chairman) (Member) (Member) CSI Communications | December 2014 | 47 Rewards & Recognitions: A reason for Double Delight There is nothing more satisfying than getting recognized for all of your eff orts in front of your peers from diff erent organisations. Recognitions received by organisations are for its ability to adapt the constantly evolving innovative norms. Such recognitions for organisations brings in double delight – one for the organisation itself as winning one only emphasizes that they have taken correct steps in the right direction and also positions itself as an enterprise that’s always at the threshold of future growth. It’s not only the organisation that gets all cheers, the people on the project also get rewarded for their explicit role in staging a successfully chalked out plan. The delight of the ‘people’ on the project forms the grounds for the second part of the delight. That’s the precise reason why we see so many awards and recognition programs across various functions in diff erent industries all around the world. Most reputed amongst them are those that are felicitated by industry associations and media houses – both being neutral bodies. Associations are much more attractive because of participation from senior industry professionals as adjudicators of the categories and the selection process. Furthermore, the status of being a non-profi t organisation adds to the credibility. I have been pre-dominantly associated with Information Technology industry and it’s easier for me to talk on how awards for technology implementation from associations have always been regarded as ‘THE RECOGNITION’ to be received. Amidst a clutter of year-on-year awards in India, ‘CSI Excellence in IT Awards’ stands out and we have seen how technology teams in the enterprise burns mid-night oil to ensure that their nominations is well structured and its increase the probability of being selected by the esteemed jury comprising. The jury for the awards comprises of Academicians and Technology leaders from the industry. The next edition for awards has been already announced and as I understand, the nominations are fast trickling in. By the time the nominations close by end of December 2014, the number of entries for various industry categories would surely be a massive chunk which may well mean loads of work for Deloitte who in the capacity of ‘Strategic Process Validator’ would scrutinise each of these papers before arriving at a shortlisted list of nominees. More information is available on www.csiawards.in. Technology companies off ering solutions to the Indian enterprise would do well to associate with such awards program in order to get the best mileage from the reach, visibility & one-on-one networking opportunities presented. I would be interested to hear about your opinion on how a recognition bestowed on you or your organisation have changed things for you. Author: Salil Warior, Director, i3R Global i3RGlobal - Event partners with CSI Excellence in IT Awards 2014 CSI Communications | December 2014 | 48 www.csi-india.org CSI Communications | December 2014 | 49 CSI News From CSI Chapters » Please check detailed news at: http://www.csi-india.org/web/guest/csic-chapters-sbs-news SPEAKER(S) TOPIC AND GIST GHAZIABAD (REGION I) Prof. RK Khandal, Puneet Garg, Dr. Kavita Saxena, RK 15 November 2014: Region I - Regional Student Convention 2014 on theme Vyas, Dr. DK lobiyal, Saurabh Agrawal, Vijay Rastogi, “Digital India-Role of Gen-Y” Anilji Garg, Dr. Arun Sharma, Dr. Rabins Porwal, Lt. Gen More than 250 students from 20 colleges from various states of region 1 Ashok Agarwal and Mr. Rajesh Dogra participated. Events such as Paperix (Paper presentation), Innovasie(Software Design, Working Model & Circuitronics), Quiz-O’-Caf@c (Technical Quiz), Posterino (Poster Paper Presentation), Bugbusters (Programming & Code debugging), Animaatio (Web Design & CAD Design), E-Warz (Gaming) & Axon (Documentary) were held. They were judged by representatives from Industry and Academia. Valedictory program Chief Guest was Lt Gen Agarwal and Guest of Honour was Rajesh Dogra. Refer to links - http://epaper.livehindustan.com/story.aspx?id=7188&boxid=63738164&ed_ date=2014-11-15&ed_code=2&ed_page=18 http://epaper.livehindustan.com/story.aspx?id=7897&boxid=70527726&ed_ date=2014-11-16&ed_code=2&ed_page=2 Guest of Honour at Valedictory NOIDA (REGION I) Anuj Agrawal, Tanmoy Chakrabarty and Gopal Krishna 24 November 2014: Seminar on “E-Governance Initiatives of GOI and Agarwal, Economic Cell, BJP Implications on Corporate Governance” Anuj Agrawal mentioned in inaugural talk that governance can be improved only with help of IT enabled citizen centric services. Keynote speaker Mr. Tanmoy said that with digitization and e-governance we can remove petty corruption. With automated processes in government services nobody will have to visit government offi ces and bribery in government offi ces will be eliminated. Gopal Krishna Agarwal said that new BJP government is diff erent in its approach towards citizen problem solving. PMO is entertaining queries and complaints online and helping to fi nd solutions. Government wants to automate services so that people can get benefi t of all Government schemes. Preparation of road map for digital Noida is to be presented to Noida authorities and constructive suggestions are being given to MSME task force of Central Government for all-round development of MSME sector. Honoring the guest RAIPUR (REGION III) Dr. Sanjay Kumar and Mr. Sisodia 7-8 November 2014: Workshop on “Android” First participants were told about basics of Android Operating System and later architecture & applications of Android followed by extensive hands on session for writing small applications like sms application, sms broadcast application to a group, event handling programming like mouse and button clicks etc. were explained. Participants were given Android Software Kit which contained Android SDK, Eclipse IDE, Java JDK and some notes. Certifi cates were distributed to participants. Practical based test on Android was also conducted in which 1st winner was Gurdeep Singh, 2nd was Devendra Ausar, 3rd was Jayshri Veram and 4th was Nidhi Jaswani. A Group photograph during Android Workshop CSI Communications | December 2014 | 50 www.csi-india.org SPEAKER(S) TOPIC AND GIST BANGALORE (REGION V) Mrs. Bhanumathi and Mr. Shamsundar Dhage 18 October 2014: Workshop on “Design Patterns” Mr. Dhage spoke on Introduction to the “Design Patterns”. Topics covered were - 1) Need and applicability of design patterns in current advanced software development 2) Categories of Design patterns and diff erent design solutions under each category 3) Understanding of various design patterns with case studies and examples 4) Exercise: Identifying and application of design pattern for given problem statements 5) Creational Patterns: to create objects in a manner suitable to the situation. Introductory Session VISAKHAPATNAM (REGION V) Y Madhusudana Rao, Prof. GSN Raju, Prof. SV Raghavan, 29 September 2014: Felicitation to Prof. SV Raghavan and technical talk on Prof. PS Avadhani, KVSS Rajeswara Rao and TNS Rao “Emerging opportunities in ICT” Chapter took honor in felicitating Prof. Raghavan by Prof. GSN Raju and Prof. PS Avadhani. Subsequently technical talk was presented by Prof. Raghavan on Emerging opportunities in ICT. In his talk he explained evaluation of big data from not only enterprises but also from non-business organizations. Present uses of Big Data in research of micro biology, nano- technology and searching techniques out of big data were also explained. L-R: Y Madhusudana Rao, GSN Raju, SV Raghavan, PS Avadhani, KVSS Rajeswara Rao, TNS Rao PUNE (REGION VI) Harry Viet, Ronald van Grunsven, HR Mohan, 10 October 2014: Conference series on “Advances in Cloud Computing Dr. Anirban Basu, Anand Joglekar, Suresh V, Atul Gore, (ACC)” Divyesh Desai, Anand Agrawal, Yogesh Kulkarni, Conference was inaugurated by HR Mohan. Harry Viet & Ronald van Grunsven Dr. Rajesh Ingle, Prafulla Wadaskar and Monish Darda. from Amsterdam, Netherlands delivered plenary keynote and covered IOT in pest control - “The use of poisons in controlling pests, results in huge damage to nature and welfare of humans and animals”. Mr. Suresh delivered tutorial on “Saving lives through IoT”, Atul Gore covered “Notes from the (greasy) fi eld - IoT Platform Components". Interesting talk on "IoT and Automobiles – Practical Applications" was presented by Divyesh Desai. Anand Agrawal covered “IoT – SAP Perspective". There was Panel Discussion on "The present and the future of IOT". Yogesh Kulkarni, Dr. Rajesh Ingle, Prafulla Wadaskar were panelists and discussion was moderated by Monish Darda. Release of proceedings CHENNAI (REGION VII) Dr. San Murugesan, Dr. Arpan Pal and Mr. H.R. Mohan 21 November 2014: Professional Development Program on “Internet of Things (IoT)” Program was organized jointly with IEEE Computer Society, Madras Chapter on “Internet of Things (IoT): Technology, Applications, and Impact - How you can capitalize on the next big thing in IT”. The PDP was inaugurated by HR Mohan who in his address highlighted that IoT is the next wave in the era of Internet and its growth is largely fuelled due to the advances in Internet, mobile, embedded devices and sensor technologies. Mr. HR Mohan inaugurati ng PDP at CSI Educati on Directorate Chennai CSI Communications | December 2014 | 51 SPEAKER(S) TOPIC AND GIST TRIVANDRUM (REGION VII) Jobin Wilson, Chitharanj Kachappilly, Prasanth 25 October 2014: Workshop on “Big Data Analytics” Sasidharan, Ranjith Uthaman, G Neelakantan and In 4 technical sessions, topics covered were (1) Introduction to Big Data Sreekanth P Krishnan (Distributed Computing, Hadoop1.x Ecosystem, Deploying a Hadoop cluster), (2) Hadoop Programming Examples (MapReduce, Hive SQL, PIG Scripts, (3) Data Sciences for Big Data (Introduction to data science Concepts, Examples, (4) Big Data Architecture and Industry use cases (How industry uses big data, How Facebook and twitter use Big Data and a Concluding session. There were 52 participants from organizations such as CDAC, VSSC, NeST, Envestnet, Triassic Solutions, Tata Elxsi, Allianz Cornhill, InApp, SCT College of Engineering, TCS, CTS, FlyTXT and MBCET. Mr. Neelakantan introducing faculti es of workshop sessions From Student Branches » (REGION-I) (REGION-I) SECOND CSI J&K STATE STUDENT CONVENTION BY MIET AMITY UNIVERSITY UTTAR PRADESH, NOIDA 26-09-14: Prof Vyas, RVP addressing the convention at MIET, Prof. M.N. 25-09-14: Glimpse of CONFLUENCE 2014 - Dr. K.M. Soni, Ms. Nitasha Hoda Division I, Sameru Sharma, Dean Engineering University of Jammu Hasteer, Dr. Abhay Bansal, Dr. Shri Kant Sharma, Dr. Ravi Prakash, and Principal GCET were present. Dr. Balvinder Shukla, Dr. Raj Kumar Buyya along with other distinguished guests. (REGION-I) (REGION-IV) G.D. GOENKA UNIVERSITY, GURGAON GANDHI INSTITUTE FOR TECHNOLOGY, BHUBANESWAR 16-10-14: Prof. (Dr.) D.P. Kothari Former Vice Chancellor, VIT University, 20-09-14: New student branch was inaugurated by Dr. Anirban Basu, Director i/c IIT Delhi, Principal, VNIT Nagpur talking about the current Division Chairman. Mr. Sanjay Mohapatra, Secretary is handing over the scenario and future scope of renewable energy in India. Academic membership certifi cate. (REGION-V) (REGION-V) GURU NANAK INSTITUTIONS TECHNICAL CAMPUS (GNITC), HYDERABAD CMR TECHNICAL CAMPUS, HYDERABAD 15-11-14: Dr. D. V. Ramana, Sr. Consultant, HP Global Soft Ltd. & Sr. CSI 13-11-2014: Workshop on “USAGE OF INFORMATICA TOOL IN Member delivered a talk on “Big Data Management, Technologies and DATAWARE HOUSING AND DATA MINING” by Mr. Sreehari Tippana, Applications” to the student branch. Wipro Technologies, Hyderabad. CSI Communications | December 2014 | 52 www.csi-india.org (REGION-V) (REGION-V) VITS COLLEGE OF ENGINEERING, VISAKHAPATNAM SUMATHI REDDY INSTITUTE OF TECHNOLOGY FOR WOMEN, WARANGAL 22-09-14: Resource person Shri Gampa Nageshwer Rao Addressing the 26-09-14: Dr. K. Rajanikanth inaugurated the CSI Student branch with audience. Dr. N. Sambasiva Rao, Principal, Dr. B. Ramasubba Reddy, CSE HOD and Mr. M. Ranjith Kumar CSI SBC. (REGION-V) (REGION-V) SRINIVASA INSTITUTE OF ENGINEERING & TECHNOLOGY, CHEYYERU JAWAHARLAL NEHRU TECHNOLOGICAL UNIVERSITY, ANANTAPUR 21-11-14: Two Days Workshop on C-Programming by Mr. K. Devarajulu 16-10-14: CSI-ED conducted a free workshop for faculty members on BOSS MOOL, jointly with CDAC, IIT Madras, and JNTU-A. The workshop, was inaugurated by Prof. Lal Kishore, Vice Chancellor, JNTU-A. (REGION-VII) (REGION-VII) AVS ENGINEERING COLLEGE, SALEM SRI RAMAKRISHNA ENGINEERING COLLEGE, COIMBATORE 25-09-14: Workshop on “Entrepreneur in you and Multimedia & Plug 28-07-14: Dr. M. Sundaresan, Chairman, CSI Coimbatore Chapter in development”. Principal Dr. G.Tholkappia Arasu, Prof. SP. Malarvizhi, inaugurated the INTERACT 2014 Dean of CSE (SBC) with student offi ce bearers. (REGION-VII) (REGION-VII) SKR ENGINEERING COLLEGE, CHENNAI EINSTEIN COLLEGE OF ENGINEERING, TIRUNELVELI 6-11-2014: FDP Workshop on “WIRELESS SENSOR NETWORKS” 07-11-2014: Motivational talk on “Key skills for Industry expectation” by Mr. A. Muruganantham, Senior manager, Hexaware Technologies CSI Communications | December 2014 | 53 (REGION-VII) (REGION-VII) NATIONAL ENGINEERING COLLEGE, KOVILPATTI SASTRA UNIVERSITY, THANJAVUR 10-10-14: National level Technical Symposium Souvenir released by Chief 4-10-14: Prof. J. Naren, SBC with Students of Final Year at the Workshop Guest Dr. R. Natarajan, Chairman, Division (II). Mr. Jerart Julus- SBC, on Python Programming Dr. Manimegalai,, Dr. Chockalingam-Director, Mr. Nirmal Lakshman- Student President (REGION-VII) RAJALAKSHMI ENGINEERING COLLEGE, CHENNAI Please send your student branch news to Education Director at [email protected]. News sent to any other email id will not be considered. Please send only 1 photo 14th & 15th November, 2014: Regional Conference on per event, not more. “Innovations in Engineering Education and Research” CSI Communications | December 2014 | 54 www.csi-india.org Application for Travel Grants for Researchers Research Committee of Computer Society of India has decided to partly fund CSI Life Members to the extent of Rs. 25000/ for travelling abroad to present research papers at conferences. CSI Life Members who have been invited to present papers abroad and have received partial or no funding are eligible to apply for the same. They have to apply within December 31, 2014 to [email protected] and furnish: 1. Name of the Applicant, Organization Details and Bio Data of Applicant 2. CSI Life Membership Number 3. Name of the International Conference with details of the organizers 4. Conference Venue and Date 5. Copy of the Research Paper 6. Copy of the Invitation Letter received from the organizers 7. Details of funding received from/applied to for funding to any other agency 8. Justifi cation for requesting support (in 100 words). 9. Two References (including one from head of the organization) Dr Anirban Basu Chairman, CSI Division V (Education and Research) CSI Communications | December 2014 | 55 Prof. Bipin V Mehta CSI Calendar Vice President, CSI & Chairman, Conf. Committee 2014 Email: [email protected] Date Event Details & Organizers Contact Information December 2014 events 12-14 Dec 2014 49th Annual Convention, Organised by Computer Society of India, Hyderabad Sri J A Chowdary Chapter In association with JNTU-Hyderabad & DRDO. Theme: Emerging ICT Sri Gautam Mahapatra for Bridging Future Venue: JNTUH, Kukatpally, Hyderabad [email protected] http://www.csihyderabad.org/csi-2014 12-14 Dec 2014 Special session on “Cyber Security and Digital Forensics” during Computer Dr. Vipin Tyagi Society of India Annual Convention - 2014 by CSI Special Interest Group on [email protected] Cyber Forensics, JNTU Hyderabad 16-20 Dec 2014 ICISS-2014: International Conference on Information Systems Security. At [email protected] Institute for Development & Research in Banking Technology (IDRBT), Hyderabad, India. Co-sponsored by CSI Division IV and CSI SIG-IS. http://www.idrbt.ac.in/ICISS_2014/ 19-21 Dec 2014 EAIT-2014: Fourth International Conference on Emerging Applications of Prof. Aditya Bagchi Information Technology at Kolkata. Organized by CSI Kolkata at Indian Statistical Dr. Debasish Jana Institute, Kolkata https://sites.google.com/site/csieait/ For paper ssubmission . Prof. Pinakpani Pal https://cmt.research.microsoft.com/EAIT2014 Prof. R T Goswami [email protected] 22-24 Dec 2014 ICHPCA-2014: International Conference on High Performance Computing and Prof. (Dr.) Rachita Misra Applications Organized by: CV Raman College of Engg. in association with CSI [email protected] Div-V and IEEE Kolkata Section http://www.ichpca-2014.in/ January 2015 events 3-4 Jan 2015 49th Annual Student Convention, Organised by Computer Society of India, Dr. DD Sarma, Hyderabad Chapter In association with GNIT, Hyderabad. Theme: “ Campus to Shri Raju Kanchibhotla Corporate” Venue: GNIT, Ibrahimpatnam, Rangareddy District Telangana. Shri Chandra Sekhar Dasaka http://www.csihyderabad.org/csi-2014 http://www.csihyderabad.org/csi-2014 http://www.csihyderabad. org/csi-2014 17-18 Jan 2015 5th National Conference on Indian Language Computing (NCILC-2015) organized Prof. (Dr.) A.K. Nayak by Department of Computer Application, Cochin University of Science & [email protected] Technology, Kochi, Kerala and CSI Division-III (Applications) [email protected] 23-24 Jan 2015 Regional Student (Region-3) organised by AESICS-CSI Student Branch, School of Prof. (Dr.) Aditya Patel Computer Studies, Ahmedabad University, Ahmedabad. Theme: “ICT for Make [email protected] In India” Prof. (Dr.) Sandeep Vasant http://www.aesics.ac.in/rsc3 [email protected] Prof. (Dr.) Kuntal Patel [email protected] 23-25 Jan 2015 National Workshop On Structural Equation Modeling Data Mining & Neural [email protected] Networks [email protected] February 2015 20 Feb 2015 First National Conference on Computational Technologies-2015 (NCCT’15) Prof. Ardhendu Mandal organised by CSI, Siliguri Chapter, Dept of Computer Science and Application, [email protected] University of North Bengal and CSI Div-V. http://www.nbucsaevents.in CSI Communications | December 2014 | 56 www.csi-india.org 26 Feb- 6 March Annual Symposium on Information Technology Research, Innovation and [email protected] Entrepreneurship Development award (ITRIED) [email protected] March 2015 11–13 Mar 2015 9th INDIACom; 2015 2nd International Conference on “Computing for Sustainable Prof. M.N.Hoda Global Development” Organized by Bharati Vidyapeeth’s Institute of Computer [email protected], Applications and Management (BVICAM), New Delhi [email protected] 21-22 Mar 2015 International Conference on ICT in Healthcare organized by Sri Aurobindo Prof. Durgesh Kumar Mishra Institute of Technology, Indore in association with CSI Indore, Udaipur Chapter [email protected] and CSI Division III and Division IV Communication. Prof. AK Nayak http://www.csi-udaipur.org/icthc-2015/ [email protected] Prof. Amit Joshi [email protected] May 2015 15–17 May 2015 International Conference on Emerging Trend in Network and Computer Prof. Dharm Singh Communication (ETNCC2015) at Department of Computer Science, School of [email protected] Computing and Informatics Polytechnic of Namibia in Association with Computer Society of India Division IV and SIG-WC http://etncc2015.org/ CSI Communications | December 2014 | 57 Student Essay Contest Harnessing the Power of ICT for our New Initiatives Computer Society of India, Chennai Chapter, in association with the IEEE Computer Society, Madras and IEEE Professional Communication Society, is pleased to announce an Essay Contest on the role of Information & Communication Technology (ICT) in India for school and college students. The contest will be in two streams: • Stream 1: Open to School Students (from 8th Standard to Plus 2) • Stream 2: Open to College Students (UG students of all disciplines) Now ICT has ubiquitous presence in India and other parts of the world, and it is being applied in various fi elds such as Manufacturing, Banking & Finance, Telecom, Healthcare, Hospitality, Transportation, Education, Agriculture, Environment, eGovernance, eCommerce, and Defence for quality and productivity improvements. India is a major force in the global IT landscape. ICT is a key driver of our economic development and accounts for about 6.5% of our GDP and provides employments to over 3 million people. The Government of India has recently launched three major initiatives – Digital India, Make in India, and Clean India, in which ICT can – and should - play signifi cant roles. Through this contest, we seek your thoughts, innovative ideas and solutions on how ICT could support and help these initiatives. We propose to share ideas from the young minds to PMO and DeitY. An eligible participant is required to submit his/her essay on any one of the following topics by 31st Dec 2014: 1 ICT for Digital India 2 ICT for Make in India 3 ICT for Clean India Submissions will be assessed by a panel of experts on criteria such as originality, novelty, applicability, potential value of the proposed idea(s) and clarity and style of presentation. The contest winners in EACH stream will be awarded the following prizes & certifi cates: One 1st Prize: Rs. 10000 • Two 2nd Prizes: Rs. 5000 of each • Four 3rd Prizes: Rs. 2500 of each • Ten Consolation Prizes: Amazon gift voucher of Rs. 1000 each • Certifi cate of Merit: For 25 short listed essays over and the above prize winning essays For more details & complete brochure, please visit the website: http://goo.gl/FziCmK For clarifi cations / queries if any, please email us at [email protected] The essay contest is supported by: Dynamic Group, Anjana Software Solutions Pvt. Ltd, HP Networking, Cognitive Platform Solutions (CPS) Pvt Ltd, Orbit Innovations and CloudReign Technologies. Please feel free to share this information to all your contacts and encourage participation in this contest. CSI Communications | December 2014 | 58 www.csi-india.org Registered with Registrar of News Papers for India - RNI 31668/78 If undelivered return to : Regd. No. MH/MR/N/222/MBI/12-14 Samruddhi Venture Park, Unit No.3, Posting Date: 10 & 11 every month. Posted at Patrika Channel Mumbai-I 4th fl oor, MIDC, Andheri (E). Mumbai-400 093 Date of Publication: 10 & 11 every month As India’s largest and one of the world’s earliest IT professional organizations, the Computer Society of India has always aimed at promoting education and research activities, especially in the advanced technological domains and emerging research fi elds. It is also committed to take the benefi ts of technological progress to the masses across India in particular to unrepresented territories. In order to promote research and innovation meeting the grass-root level ICT needs and emphasize the importance of joint research by faculty-students, the CSI has been providing R&D funding for last several years. The CSI Student Branches and member institutions are requested to motivate the young faculty members and students (including undergraduate and postgraduate) to benefi t from this scheme. The proposals (based on the ongoing or new projects for the academic year 2014-2015) with the following aims/objectives, expected outcome, indicative thrust areas for research funding may be submitted to: The Director (Education), Computer Society of India, Education Directorate, CIT Campus, IV Cross Road, Taramani, Chennai 600113. Last date for Receipt of Proposals: 31st January 2015. Aims and Objectives • To provide fi nancial support for research by faculty members, especially for developing innovative techniques and systems to improve teaching-learning and institutional management processes. • To provide fi nancial support to students for developing new systems catering to the needs of socially relevant sectors and/or involving proof of concepts related to emerging technologies • To facilitate interaction/collaboration among academicians, practitioners and students • To develop confi dence and core competence among faculty/students through research projects • To foster an ambience of ‘Learning by Doing’ and explore opportunities of industry funding and mentoring for inculcating professionalism and best practices among students and faculty • To recognize innovation and present excellence awards for path-breaking projects through CSI YITP awards and industry associations, Govt. Agencies and professional societies. Expected Outcome • Identifi cation of thrust areas, capability assessment, gap analysis, recommendations and future education and research directions • Integration of research methodologies into the university teaching-learning process and evolving a quality control mechanism for academic programmes and curricula • Strengthening of industry-institutes interaction through commercialization of technologies and products developed by students and faculty • Publication of research studies (ICT penetration, technological innovation, diff usion & adaptation), state-of-the-art reports and case studies of education/ research initiatives • Identifi cation of potential new and innovative projects of young faculty, researchers and students for possible business incubation Indicative Thrust Areas for Research funding The fi nancial assistance up to Rs 50,000/- for hardware projects and up to Rs 30,000/- for software projects would be provided to cover items like equipment, books/journals, fi eld work, questionnaire, computation work and report writing. The indicative thrust areas for funding include (but not limited): Technology- OS, Programming Languages, DBMS, Computer & Communication Networks, Software Engineering, Multimedia & Internet Technologies, Hardware & Embedded Systems Process & Tools- Requirements Engineering, Estimation & Project Planning, Prototyping, Architecture & Design, Development, Testing & Debugging, Verifi cation & Validation, Maintenance & Enhancement, Change Management, Confi guration Management, Project Management, Software Quality Assurance & Process Improvement, Vertical Applications- Scientifi c Applications, Enterprise Systems, Governance, Judiciary & Law Enforcement, Manufacturing, Healthcare, Education, Infrastructure, Transport, Energy, Defence, Aerospace, Automotive, Telecom, Agriculture & Forest Management, Inter-disciplinary Applications- CAD/CAM/CAE, ERP/SCM, EDA, Geo-informatics, Bioinformatics, Industrial Automation, CTI and Convergence. Last date for Receipt of Proposals: 31st January 2015 Further details and application form can be downloaded from the link of “Student’s Corner - CSI Education Directorate” at www.csi-india.org Director (Education) Computer Society of India Education Directorate CIT Campus, IV Cross Road Taramani, Chennai-600113 CSI Communications | December 2014 | 59