Second Phase of Openlab Begins

CCEJunCompNews18-21 16/5/06 12:45 Page 18

COMPUTING NEWS Compiled by Hannelore Hämmerle and Nicole Crémel COLLABORATION Second phase of openlab begins

The second phase of the CERN openlab was inside the Grid. officially launched with a ceremony at CERN The Grid Interoperability Centre is proposed on 17 May. The industry partners in the as a reinforcement of the second phase of the second phase are HP, Intel and Oracle. It will Enabling Grids for E-science (EGEE) project build on experience from the first three-year led by CERN, allowing the CERN openlab phase, during which time the partnership partners to take part in integrating and between CERN and leading IT companies certificating Grid middleware. The project will produced many excellent technical results. focus on three activities: testing and The flagship project of the first phase of certification of the EGEE middleware stacks openlab was the CERN opencluster, an on testbeds provided by the partners; advanced computing and storage cluster, support, analysis, debugging and problem which contributed to several landmark results resolution to deal with the problems in CERN’s Grid and high-speed networking encountered on the contributed testbed; and activities. The partners during the first phase interoperability efforts that review current were Enterasys, HP, IBM, Intel and Oracle. The levels of Grid interoperability, also with opencluster results showed that CERN middleware stacks proposed by the partners. openlab is a novel and effective framework for In addition to these centres, CERN openlab collaboration between multiple industry has launched an initiative in computer security. partners, in a pre-competitive spirit and based Jarno Laitinen, a CERN openlab student, Initially, the bulk of this effort will be in malware on open standards. Activities for the start-up of checks the routers for the CERN opencluster. protection, anti-spyware, intrusion detection the second phase of CERN openlab are based and intrusion prevention, with a particular focus around a platform competence centre, a Grid environment presented by a virtual-machine on client security and mail-server security. The interoperability centre and IT security activities. hypervisor, independent of all the hardware Finnish companies F-Secure and Stonesoft will The Platform Competence Centre focuses intricacies. Software and hardware join CERN openlab as contributors, a status on platform virtualization, as well as software optimization is a vital part of the deployment created to allow smaller IT companies to and hardware optimization. Platform of the Large Hadron Collider Computing Grid, participate on targeted topics for a shorter virtualization enables Grid applications to as the demand for resources by researchers is period. The Helsinki Institute of Physics was benefit from a highly secure and standardized likely to outstrip available resources, even instrumental in setting up these activities.

SIMULATION STUDIES including elementary-particle physics and chromodynamics (QCD), the fundamental nuclear physics. The new supercomputer theory of strong interactions. By simulating KEK installs Japan’s system, which started up on 1 March, consists QCD on a computer, the researchers expect to of the Hitachi SR11000 model K1 with a peak develop their understanding of the masses fastest simulation performance of 2.15 × 1012 floating point and reactions of particles such as protons, operations a second (2.15 Tflops) and the IBM neutrons and mesons (CERN Courier June supercomputer Blue Gene Solution with a peak performance 2004 p23). The supercomputer system at of 57.3 TFlops. At the time of installation, this KEK is shared with researchers from other The High Energy Accelerator Research system constituted the highest computing Japanese laboratories and universities Organization, KEK, in Japan has installed a power in Japan. working on particle- and nuclear-physics new supercomputer system for simulation The supercomputer is used mainly for the simulations, as well as those working on the studies of high-energy accelerator science, lattice simulation of quantum accelerator facilities at KEK.

Les gros titres de l’actualité informatique L’openlab entame sa deuxième phase 18 Le bulletin CNL fête ses 40 ans 20 Le KEK installe le superordinateur de simulation le plus rapide Des nanopoints magnétiques à Brookhaven 20 du Japon 18 Le champion du monde des réseaux de fibres optiques 20 Un forum EGEE rassemble des utilisateurs de la Grille 19 Plus d’autonomie grâce à un nouveau clavier braille 21 Produits 19 Evénements 21

18 CERN Courier June 2006 CCEJunCompNews18-21 17/5/06 10:50 am Page 19

COMPUTING NEWS

EVENTS EGEE forum unites Grid users

One of the plenary sessions in full swing at the first EGEE User Forum, held at CERN in March.

The first User Forum of the Enabling Grids for quantum chemistry were just a few of the E-science (EGEE) project took place on applications discussed. Digital radio 1–3 March at CERN. The event allowed broadcasting and financial modelling, as well members of the expanding community of Grid as research from outside Europe, such as users to meet to compare experiences and Taiwan’s National Digital Archive Project, were strengthen ties, and it enabled leaders of the also highlighted. EGEE project to update users on the status The forum mixed plenary sessions, and recent developments within the project describing the EGEE infrastructure and the and the infrastructure. achievements of several application More than 90 abstracts from new and communities, with two afternoons of parallel established applications were considered for sessions. The first set of parallel sessions was presentation and the some 250 attendees organized around application domains, the considerably surpassed the target of 150. The second was thematic with presentations abstracts submitted showed the variety of grouped by different issues in Grid computing. Grid applications, from the more standard in This User Forum was the first in a series, biology and high-energy physics, for example, with the next event planed for spring 2007. to others in more unusual fields, such as The next major EGEE event will be the EGEE’06 online gaming and archaeology. conference on 25–29 September in Geneva. Scientists at the meeting shared their ● The full programme and all presentations of research results and plans through the User Forum are available online, together presentations and posters. Protein with a booklet containing all 90 abstracts sequencing, medical imaging, fusion-energy submitted to the event. For details see research, gamma-ray astronomy, climate http://indico.cern.ch/conferenceTimeTable. modelling, earthquake simulations and py?confId=286.

PRODUCT INFORMATION simplify the XML data-integration tasks faced by developers. The significant performance and Stylus Studio, the industry-leading provider of extensibility enhancements to XML-deployment XML development tools and components for adapter technologies provide a compelling, advanced data integration, has announced the standards-based solution for working with availability of Stylus Studio 2006 Release 2 heterogeneous data sources. For a free trial XML Enterprise Edition. The new release download go to www.stylusstudio.com/ provides powerful new XML tools and utilities to xml_download.html.

CERN Courier June 2006 19 CCEJunCompNews18-21 17/5/06 10:43 Page 20

COMPUTING NEWS

CNL ARCHIVE electronic mail has to be paid for!), but also the superstitious nonsense in chain letters CNL is 40 years old about “bad luck if you break the chain” might actually worry some people. If you receive After CERN’s 50th anniversary in 2004, this such a letter, please have the common sense year there is another (admittedly smaller) to break the chain and send an appropriate anniversary: the first CERN Computer comment to the person who wasted your time Newsletter (CNL) was circulated in 1966. As by sending it. (CNL 188 March–May 1987.) CNL celebrates its 40th year, we will take a look back at some of the highlights – and Removal of central card punching and historical curiosities – of the early years of card-reading services computing at CERN, seen through its pages. The last central card punch and card reader, You will find a brief look at previous issues of presently connected to the CDC 835, will not the CNL here; for an extended retrospection be available after the Christmas–New Year read the CNL, which is also available online shutdown. (CNL 171, Nov–Dec 1983.) from www.cerncourier.com. View of part of the Computer Centre in 1988. The CRAY X-MP/48 supercomputer can be CRAY introduction WARNING: Future electronic-mail address seen in the background against the rear wall. At the June meeting of the Finance Committee, change for CERN the acquisition of a CRAY X/MP-48 computer In order to conform to new de facto [email protected] or [email protected]. was agreed, for installation in October 1987. standards, and to prepare for transition to CERN). (CNL 193 Sept–Dec 1988.) The system has 4 processors with a basic international standards, we are planning to clock period of 9.5 ns, 8 million 64-bit words introduce new electronic mail addresses for Computerized foolishness of bipolar main memory organized in 32 CERN users, namely [email protected] or, An electronic-mail chain letter has recently banks with a 38 ns cycle time, and for users of VMS Mail, [email protected]. been intercepted at CERN. Not only is such 128 million 64-bit words of secondary CERN.CH […] (currently addresses are foolishness a waste of time and money (yes, memory. (CNL 185 May–Sept 1986.)

STORAGE MEDIA HIGH-SPEED LINKS Brookhaven creates magnetic nano-dots Fibre-optic network sets world record

In March, scientists from a MultiTeraNet project, funded by the German Federal Ministry 2000 nm 500 nm of Education and Research, established a new An individual magnetic element showing local magnetization under an applied field via world record with colleagues from Fujitsu. They magnetic phase imaging and various reconstruction methods. From left to right: Lorentz transmitted data at 2.56 Tbit/s over 160 km – contrast; phase image; phase contours; colour vector map; and section of colour vector map equivalent to 60 DVDs a second. The fastest showing cross-tie structure. (Courtesy Yimei Zhu, Brookhaven National Laboratory.) high-speed links currently carry data at a maximum 40 Gbit/s, around 50 times slower. Researchers from the US Department of building blocks for new nanoscale magneto- The Berlin-based group smashed the existing Energy’s Brookhaven National Laboratory electronic devices and data-storage media. record of 1.28 Tbit/s, held by Japanese presented new work assessing the properties Using a state-of-the-art, field-emission researchers, which had stood for five years. of materials that may lead to magneto- transmission electron microscope equipped To enable telecommunications networks to electronic devices on the scale of billionths of with a custom-made objective lens, the group cope with the massive surge in data traffic a metre at the March American Physical can probe the magnetic properties (including from the Internet community, researchers are Society meeting in Baltimore, Maryland. spin orientation) of each dot, and map how focusing on new systems to increase data Yimei Zhu and his group at Brookhaven have the spins flip in response to an external transmission rates. In fibre-optic cables, data fabricated patterned magnetic films by magnetic field – or other variables such as are transmitted using ultrashort pulses of light depositing magnetic materials such as temperature, environment and crystal defects. and are normally encoded by switching the Permalloy and cobalt in patterns of dots, The technique uses an extremely coherent laser on and off to represent binary data. The squares or ellipses across a surface of source of electrons to produce high-resolution researchers have squeezed more data into nonmagnetic substrates such as carbon or images of unprecedented quality in which the each pulse by packing four, instead of the silicon nitride. With each dot measuring about amplitude and direction of local previous two, binary data states in a light 100 nm across, these materials could serve as magnetization can clearly be seen. pulse using phase modulation.

20 CERN Courier June 2006 CCEJunCompNews18-21 16/5/06 12:46 Page 21

COMPUTING NEWS

INNOVATION New Braille keyboard extends independence

Researchers from the Universitat Autònoma de Barcelona (UAB) and Spain’s Organization of the Blind (ONCE) have developed a Braille keyboard for PCs that provides more applications for blind people and is particularly useful for scientific texts and musical scores. Until now, this kind of task required assistance from a sighted person. The new keyboard connects to the PC through the USB port. It will make using a computer much easier for blind people who are accustomed to using Braille typewriters UAB professor Jordi Roig with the new The new Braille keyboard has eight Braille because this is the first Braille keyboard to Braille keyboard developed in Spain. keys, which allow typing in any language. combine the function and movement keys of a conventional keyboard with eight Braille keys The project, proposed by ONCE, began in the research. The group is currently working on that allow the user to write in any language. March 2004 and was coordinated by UAB other projects that will enable blind people to Users can also select between a cumulative Department of Microelectronics and Electronic work autonomously with technological and a corrective writing mode. In the Systems researcher Jordi Roig, who has been applications. These projects include cumulative mode, the dots of a Braille symbol visually impaired for three years and is a developing an automatic conversion tool for are indicated by pressing the keys one after member of ONCE. Teachers from the UAB websites that are not currently accessible to another; in the corrective mode, this is done by School of Engineering and the UAB School of blind people and creating touchable screens pressing the necessary keys simultaneously. Computer Science in Sabadell participated in that raise the information so that it can be felt.

Calendar of events June Computing for Computational Science on Parallel Computing (Euro-Par 2006) 18–21 PARA’06 Rio de Janiero, Brazil, Dresden, Germany, www.europar2006.de/ Umeå, Sweden, http://vecpar.fe.up.pt/2006/ www.hpc2n.umu.se/para06/ September 15–19 2nd Euroscience Open Forum 13–15 HPCC-06 19–23 15th IEEE Symposium on High- (ESOF2006) Munich, Germany, Performance Distributed Computing Munich, Germany, www.esof2006.org http://hpcc06.lrr.in.tum.de/ (HPDC-15) Paris, France, www.hpdc.org/ 16–18 GRIDNS 2006 25–29 EGEE’06 Silicon Valley, California, US, Geneva, Switzerland, 27–30 21st International Supercomputer www.iaria.org/conferences/ICNS06.html www.eu-egee.org/egee06 Conference (ISC2006) Dresden, Germany, www.supercomp.de/ 17–21 22nd Asia–Paciﬁc Advanced 25–28 Cluster 2006 Network (APAN) meeting Barcelona, Spain, http://cluster2006.org July National University of Singapore, Singapore, 3–5 EuroPython conference www.apan.net/meetings/future.htm 28–29 The 7th IEEE/ACM International CERN, www.europython.org Conference on Grid Computing August (Grid 2006) 4–7 26th International Conference on 1–4 IFIP International Conference on Barcelona, Spain, www.grid2006.org Distributed Computing Systems (ICDCS Embedded And Ubiquitous Computing 2006) (EUC’2006) December Lisbon, Portugal, Seoul, Korea, http://.euc06.euc- 4–6 e-Science 2006 http://icdcs2006.di.fc.ul.pt/ conference.org/ Amsterdam, Netherlands, www.escience-meeting.org/eScience2006/ 10–12 VECPAR’06: High Performance 29–1 September European Conference Paper deadline: 1 August 2006

CERN Courier June 2006 21 AdvertorialOneCol_procurv_Final 24/1/06 2:24 pm Page 1

ADVERTISING FEATURE Intelligent networking – a vital part of today’s IT infrastructure

Whether you consider a computer in its tra- By Wenceslao Lada offer free software upgrades to enable our ditional sense to be a means of calculating Vice President and General installed base of equipment to take advan- large numbers quickly or your understand- tage of new networking protocols as they ing is closer to the French word for a com- Manager, ProCurve Networking, become available. puter – ordinateur – which implies a tool Europe, Middle East & Africa Today’s networks are no longer just for organising and managing work, it is about connecting computers in a server clear in this day and age, such a device has a room to computers on a desktop. As limited use unless connected to other com- employees become more mobile, they want puters via a reliable and intelligent network. to access their corporate networks from This is as true for a nuclear research notebook computers, PDAs or even the lat- establishment as it is for a commercial est generation of mobile phone. They will office although the amount of computing be calling in from wherever they happen to power available to each may be vastly dif- be, which may be from a hotel room, an air- ferent. Each organisation needs its employ- port or an Internet café. The modern net- ees to be able to communicate quickly and work has to be able to handle these requests easily with each other and the outside from a variety of media and to respond world; it needs to be able to add and remove without sacrificing ease of use or security. users from the network with the minimum As such, the need to devise policies cen- of fuss and effort, and it needs to be able to trally is as important as ever, so that stan- guarantee its security and integrity. dard operating procedures can be As a key supplier of networking equip- developed and maintained, but the neces- ment to CERN, ProCurve Networking by HP sity to devolve the implementation of such has had to ensure that the networks we sup- underpins our whole product design and policies to the edge of a network is also ply are able to deliver on all the require- the intention is to allow control of policies vital. Otherwise, traffic to the centre ments of a world-class organisation. That we regarding security and resource allocation would become so congested as to be are able to do so owes a lot to our own com- to be decided upon at the centre of the net- unworkable and the likelihood of employ- mitment to research and development and work while allowing implementation to be ees attempting to circumvent such policies careful attention to the most important part pushed out to PC users at the edge. To do would increase. Far better for employees to of a modern network – the ability to man- this we have had to build in significant intel- have a workable system that they can trust age it so that control is firmly in the hands ligence to all our products ranging from the than to have a well-meaning but autocratic of management but the ability of the indi- large complex switches and routers which and totally unworkable system that they vidual users to operate it is not impaired. sit in the centre of the network to intelligent will lose faith in. This conundrum is not as easy to solve switches which are to be found at the edge. ProCurve Networking by HP has over 25 years as it may appear. Traditionally, network All ProCurve’s networking products and experience in the networking industry and has security could be realised quite simply by solutions are based on proven open indus- been responsible for many innovations. Its concentrating all decision making in the try standards such as Ethernet switching, customers include some of the world’s leading hands of those controlling the central core 802.11 wireless networking and other companies such as SAP, Hilton Hotels and of the network and imposing great restric- open standards for delivering power to Kodak. Its products are also used in leading tions on access, bandwidth and flexibility network devices via the network connec- universities and research establishments to users throughout the organisation. tion itself, the so-called Power over Ether- throughout Europe including Sahlgrenska in An alternative, to have loose central con- net standard. Adhering to open industry Sweden, the British Library and the University trol and devolve responsibility for access standards means we are able to offer our of York in the UK. control, security and bandwidth control to customers best in class value as well as departments at the edge of the network maximum choice and flexibility. Further information on ProCurve Networking gave great freedom to users but left the way Furthermore, the majority of ProCurve solutions and products is available at: open for all manner of abuse – ranging Networking products and solutions are www.hp.com/eur/procurve from unauthorised access to company data based on our own custom designed ASICs to inefficient use of network infrastructure. (application specific integrated circuits) which capture our design expertise in Adaptive EDGE Architecture™ mass-produced silicon chips. By cutting ProCurve Networking by HP has met this down the number of components needed, challenge with a strategy we call the Adap- these ASICs allow us to offer life-time guar- tive EDGE Architecture™ or AEA. This antees on many of our products and to CCEJunCHEP23-25 15/5/06 16:00 Page 23

CHEP06 Computing conference in India gets ready for LHC

Sunanda Banerjee reports from the CHEP06 conference held this year in Mumbai, India.

Computing in High Energy and Nuclear Physics (CHEP) is a major series of conferences that has been held at roughly 18-month inter- vals since 1985, alternating between Europe, North America and other parts of the world. The Tata Institute of Fundamental Research (TIFR), Mumbai, India, organized the 15th in the series, CHEP06, on 13–17 February. The main theme of the conference was to review progress in making the Grid a powerful and reliable computing resource in time to process data from the Large Hadron Collider (LHC). This was the reason for a three-day workshop on the 4th Service Challenge in the Worldwide LHC Computing Grid (WLCG) that pre- ceded the conference. The conference also aimed to learn from the experience of experiments that are currently running, to stay in touch with Grid applications in other sciences and to look to the future. The conference opened with a note that this may be the last conference in this series before LHC start-up. Consequently the emphasis on the readiness of the LHC accelerator, experiments and computing services was of major interest. In the opening talk Jos Engelen, chief scientific officer at CERN, confirmed that the LHC project, the machine, detectors, and the WLCG are well underway for physics in 2007. Schedules are tight, but not impossible to meet. There is a large potential for exciting physics.

Grid computing Jamie Shiers of CERN elaborated on the readiness of the LHC computing facilities. There has been excellent progress on the three key areas addressed by the WLCG, namely data-transfer tests, service availability and time to resolve problems, and provisioning of resources. A clear plan has been established for the remaining work that is needed to meet the computing challenges of the LHC, and The presence of the President of India, A P J Abdul Kalam (left), success can be achieved through a pragmatic, focused and co-oper- showed the importance of high-level IT services for India and ative approach. proved to be a main attraction at this year’s CHEP conference. Focusing on WLCG services, Les Robertson of CERN explained where expectations have been fulfilled and where more efforts are sented Grid activities in Japan and China. The primary objective of needed to be ready for the first LHC beams. Two Grid infrastruc- the National Research Grid Initiative project in Japan is a seamless tures, LCG and the Open Science Grid (OSG), are now in operation federation of heterogeneous resources. It started with computations for the computing services for the LHC and the main priority is to set in nano-science and Grid technology and expanded its scope to up a reliable operation of baseline services. Ruth Pordes from OSG high-energy physics, astronomy and biology. The Grid activities in emphasized the interoperability aspects of different Grids in view of China encompass a variety of disciplines from high-energy physics the various Grid infrastructures that exist around the world today, and bioinformatics to environmental science, material science, com- and the many ongoing efforts in this field. putational chemistry and others, which are driven by experiences Kenichi Miura of the National Institute of Informatics in Japan and from LCG. Grid activities in the Asia Pacific Federation span many Gang Chen of the Institute of High Energy Physics in Beijing pre- fields, from high-energy physics and atmospheric science to nano- ▲

CERN Courier June 2006 23 CCEJunCHEP23-25 15/5/06 16:00 Page 24

CHEP06 The PHENIX experiment at the Relativistic Heavy Ion Collider at Brookhaven has data rates similar to the ones expected from the LHC. Martin Purschke of Brookhaven talked about the experiences from the experiment, which was originally designed for a data rate of 20 MB/s, but eventually operated 60 times faster. Data compres- sion and buffering were key ingredients in dealing with these amounts of data. Peter Elmer of Princeton University discussed distributed data-management issues in high-energy physics experiments in general. Gaining from experience in the ongoing experiments, a great deal of effort is put into understanding the data access pat- tern in the next generation of LHC experiments. Elmer foresees much exciting work related to data management in the next 3–5 years. Elizabeth Sexton-Kennedy of Fermilab talked about the social and technical challenges of the event-processing framework. She concluded that frameworks should be judged by how well they meet the technical challenges of complexity and scalability. In addition, frameworks provide a set of tools that organize a large group of developers into moving in the same direction.

Anil Kakodkar, secretary to the Department of Atomic Energy in Networking India, spoke about the importance of CERN–India collaboration Networks used by high-energy physics and other data-intensive sci- in the LCG project at the opening session of the conference. ences are advancing rapidly, as shown by Harvey Newman of CalTech. These fields are also learning to use long-range networks and biomedical applications, as pointed out by Simon Lin from more effectively. Hybrid dark fibres are emerging as the means for Academica Sinica in Taipei. rapidly increasing communication speed in many countries. He Piergiorgio Cerello of the Istituto Nazionale di Fisica Nucleare emphasized that it is now important to close the digital divide to showed the role of the Grid in medical applications, and Rajiv Gavai allow scientists in all world regions to take part in discoveries. of TIFR, addressed computing challenges in lattice quantum Focusing on connectivity issues in India, Ashok Jhunjhunwala of the chromodynamics. Mathai Joseph of the Tata Research Development Indian Institute of Technology, Madras, showed that Internet tech- & Design Centre, elaborated on the software development process nology will have a huge impact on the lives of rural communities in and computing challenges for astronomy experiments, such as the developing countries if there is a big enough vision. In India, there Square Kilometre Array. should be 50 million broadband connections by 2010 and this will double the per capita GDP in rural India. Many innovations are Software and data taking needed to meet this effort and, in particular, the need for an unin- Paris Sphicas of CERN focused on the start-up of LHC data taking terrupted power supply will cause a bottleneck. and the state of readiness of the software for LHC experiments. Most of the common software is in place and much of the experiment- Industrial contributions specific software is either complete or has a fully functional proto- There were several exciting presentations from IT industries from all type. Important tests such as calibration or Grid utilization are over the world. Tony Hey of Microsoft talked about the capability to underway, deployment has begun in earnest and performance stud- access, move, manipulate and mine data in collaborative science ies are in progress. There is still a long way to go to get some of the applications. He emphasized the continuing trend towards decen- complicated analyses carried out, but all of the experiments are tralized, networked resources and using the Internet and open approaching that goal. access for promoting the global scientific knowledge base. Rene Brun of CERN presented the successful ROOT project, an Alan Gara of the IBM T J Watson Research Center discussed the object-oriented framework for large-scale data analysis, in particular role of high-performance computing in a variety of physics applica- its adaptation in the era of the multi-core CPU. Instead of pushing tions. He also talked about scalability issues in supercomputing gigabytes of source or shared libraries to Grid working nodes, he and emphasized that architectural innovation is critical to continuing proposed a pull technique, where only the software necessary to performance scaling. David Axmark of MySQL AB explained how to run an application is downloaded in an incremental way. While the make a scalable database system. In favour of free software, he core ROOT software is being consolidated to be ready for LHC data spoke of software patents as a threat to software innovation. Free taking, prototyping work is underway for the new concept. access to the source code means that the database system Beat Jost of CERN discussed design criteria and the resulting becomes more robust, as well as secured. architecture of future data-acquisition systems, which have to Google’s approach to organizing the entire world’s information and change owing to a move towards eliminating hardware triggers. For making it universally accessible and useful was explained by Lalitesh more flexible triggering and optimal efficiency, the readout will be at Kathragadda of Google India. He explained their in-house solution bunch- or train-crossing points at the detectors and data will be sent for data storage, for running jobs on pools of machines, and for sim- out to CPU farms for event selection. plifying large-scale data processing. Anirban Chakrabarti from Infosys

24 CERN Courier June 2006 CCEJunCHEP23-25 17/5/06 10:44 Page 25

CHEP06 Events around CHEP06 The three-day pre-conference workshop on LCG service challenges, organized by Jamie Shiers of CERN, attracted 155 participants. It was devoted to the primary activities of the programme for the WLCG Service Challenge in the run-up to the full WLCG production service due to commence in October. The topics concerned were data and storage management, WLCG services in general and the experiments’ computing models and plans to validate their offline frameworks using the WLCG services during 2006. The workshop was highly interactive, with many profitable discussions and breakout sessions. The first day of the workshop was spent on storage management, focusing on questions of interoperability between different implementations of storage resource managers, as well as deployment issues. Key issues that were discussed during the second day were the service-level targets defined in the WLCG Memorandum of Understanding, how they could be met and how they would be measured. The third day was spent discussing the impact of the Participants at CHEP06 read one of the 144 posters displayed computing models of the different LHC experiments on the in the foyer of the Tata Institute of Fundamental Research. service requirements at a multi-disciplinary Tier-1 or Tier-2 site. In the two days before the conference, François Fluckiger of Technologies, a leading software company in India, elaborated on CERN organized a tutorial programme with six lecture series on various themes of Grid applications from middleware to specific appli- the fundamentals of Grid technology, cryptography and security cations where Infosys is focusing its current research activities. protocols, fundamentals of quality of service in networking, security from theory to implementation, pragmatic software Key attractions engineering and databases. The 55 participants found the A key attraction at the conference was the visit of A P J Abdul Kalam, experience extremely positive. the Indian president, who gave the valedictory address. He summa- A one-day workshop, organized jointly by TIFR and CDAC, rized the missions for CHEP06: computing in particle physics, space followed the conference and focused on ways to improve and particle research, and energy. He encouraged a collaboration connectivity issues in India. Delegates from the Knowledge between CERN and the Indian Department of Atomic Energy on Grid Commission in India and the government ministries on activities and emphasized that Grid activities should be enlarged to information technology attended the workshop. Internet-2, CDAC encompass the Knowledge Grid, the e-Governance Grid and the and the Education and Research Network signed a Memorandum Health-Care Grid, and interconnect with the Providing Urban ameni- of Understanding to promote high-bandwidth connectivity in India. ties in Rural Areas Grid that connects a billion people across India. Another attraction of CHEP06 was the setting up of a 622 MB/s link to the US through Japan, over which several application jobs were Further reading demonstrated. This was achieved through a collaboration between For further information see www.tifr.res.in/~chep06/. CalTech, Internet-2, the World Bank and the Centre for Development of Advanced Computing (CDAC) in India. This effort drew the attention Résumé of the funding authorities of India who appreciated the role of the Conférence sur l’informatique à Mumbai high-bandwidth network for research and education. The final session of CHEP06 was a panel discussion to address the Entre le 13 et le 17 février s’est tenue à l’Institut Tata de importance of bridging the digital divide between different countries. recherche fondamentale la conférence CHEP06, consacrée à The panellists, S Ramakrishnan (CDAC, Pune), Harvey Newman l’informatique en physique nucléaire et des hautes énergies. Elle (CalTech), Viatcheslav Ilin (Moscow State University), Alberto Santoro a fait le point sur l’avancement de la Grille de calcul, une (Universidade do Estado do Rio de Janeiro) and A S Kolalskar ressource informatique puissante et fiable qui devra être prête à (University of Pune), tried to formulate possible ways to address this temps pour traiter les données du Grand collisionneur de hadrons problem for nations where low bandwidth connectivity still prevails. (LHC). La conférence visait également à tirer les enseignements Co-conveners Sunanda Banerjee and Atul Gurtu of TIFR concluded des expériences en cours, à suivre les évolutions des applications the conference by thanking the 478 participants, the more than 250 de grilles dans d’autres domaines scientifiques et à se tourner speakers and chairs in the plenary and the eight parallel sessions, vers l’avenir. Elle a été précédée d’un atelier de trois jours sur le the members of the International Advisory and Programme commit- quatrième essai d’exploitation de la Grille mondiale de calcul du tees, as well as the funding agencies, sponsors and all those who LHC (WLCG). worked hard to make the conference a success. The next conference in the series will be held in Victoria, Canada, in September 2007. Sunanda Banerjee, TIFR.

CERN Courier June 2006 25 Accelerating Technology

ASP Injection System commissioning according to schedule

Danfysik was awarded the contract to deliver the 3GeV Injector turnkey system to the Australian Synchrotron Project in Melbourne. Within the contractual time frame, 27 months after project start, the Injection System was running at 3 GeV, sending electrons into the Booster to Storage Ring Transfer Line.