DOCSLIB.ORG

Lodflow – a Workflow Management System for Linked Data Processing

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Lodflow – a Workflow Management System for Linked Data Processing LODFlow – a Workflow Management System for Linked Data Processing Sandro Rautenberg Ivan Ermilov Midwestern State University AKSW/BIS (UNICENTRO) University of Leipzig Guarapuava, PR, Brazil Leipzig, Germany [email protected] [email protected] leipzig.de Edgard Marx Sören Auer Axel-Cyrille N. Ngomo AKSW/BIS University of Bonn / AKSW/BIS University of Leipzig Fraunhofer IAIS, Germany University of Leipzig Leipzig, Germany [email protected] Leipzig, Germany [email protected] [email protected] leipzig.de leipzig.de ABSTRACT 1. INTRODUCTION The extraction and maintenance of Linked Data datasets is The creation and maintenance of Linked Data datasets is a cumbersome, time-consuming and resource-intensive ac- currently performed by extensive manual workflows or propri- tivity. The cost for producing Linked Data can be reduced etary scripts, which are not portable, reusable, and difficult by a workflow management system, which describes plans to maintain. For Linked Data extraction and publication to systematically support the lifecycle of RDF datasets. We projects such as LinkedGeoData or DBpedia, a long number present the LODFlow Linked Data Workflow Management of activities is required in order to produce a particular ver- System, which provides an environment for planning, execut- sion of the datasets. These activities include, for example, ing, reusing, and documenting Linked Data workflows. The downloading and loading dumps of the raw data, running LODFlow approach is based on a comprehensive knowledge various extraction tools, performing quality analysis and model for describing the workflows and a workflow execution sanity checks, loading the resulting data into a SPARQL end- engine supporting systematic workflow execution, reporting, point for publication, generating links to other datasets etc. and exception handling. The environment was evaluated in Performing such Linked Data production and maintenance a large-scale real-world use case. As result, LODFlow sup- activities requires in-depth knowledge of scripting languages, ports Linked Data engineers to systematically plan, execute various RDF serialization and publication technologies, tools and assess Linked Data production and maintenance work- such as SILK, LIMES for linking, and standards such as flows, thus improving efficiency, ease-of-use, reproducibility, R2RML for relational data mapping. As a result, such pro- reuseability and provenance. cesses are cumbersome, time-consuming, and error-prone, The environment was evaluated in a large-scale real-world use requiring substantial skills. case. As result, LODFlow supports Linked Data engineers to systematically plan, execute and assess Linked Data produc- In other domains, planning, executing, and orchestrating tion and maintenance workflows, thus improving efficiency, activities is supported by workflow management techniques ease-of-use, reproducibility, reuseability and provenance. and systems. With the Business Process Execution Lan- guage (BPEL), for example, a comprehensive language for Categories and Subject Descriptors the planning, execution, and orchestration of Web Services exists. There is also a large number of products from vendors H.4.1 [Office Automation]: Workflow management including SAP, IBM, and Apache supporting BPEL available on the market. In addition, for scientific workflows, a num- General Terms ber of languages and tools supporting the workflow design Linked Data, Workflow Management, Linked Data Workflow and execution exist. Specialized scientific workflow systems, Management System. e.g. Discovery Net [3], Apache Taverna [9] and Kepler [12], provide a visual programming front-end enabling user to construct their applications as a visual graph by connecting nodes together, and tools have been developed to execute such workflows in a platform-independent manner [10]. In this article, we present a comprehensive approach for defining, planning, orchestrating, and executing Linked Data production and maintenance workflows, the Linked Open Data workFlow system (LODFlow). LODFlow is based on defining workflows comprising reusable activities and steps where it is defined by Workflow Management Coalition as using a comprehensive ontological model, the Linked Data \the automation of a business process, in whole or part, during Workflow Project Ontology (LDWPO). We implemented a which documents, information or tasks are passed from one workflow execution environment, which takes workflows de- participant to another for action, according to a set of proce- fined in LDWPO ontology as an input, executes the various dural rules" [18]. Essentially, a Workflow orchestrates steps steps defined in the ontology, and generates detailed reports that transform resources in order to reach a desired result. about the execution and potential errors in machine and Benefits of systematic workflow management include explicit human-readable ways. The LODFlow is able to launch a description of activities, reproducibility, automatization and number of specialized tools such as DBpedia DIEF or Spar- generally increased efficiency. qlMap for extraction and mapping, SILK for linking, Luzzu for quality analysis. The LODFlow preserves provenance According to [7], Workflow Management is a technology sup- and adds comprehensive metadata, such as the version, invo- porting the reengineering of business and information pro- cation, and configuration of the tool execution in a concrete cesses. The main components of the Workflow Management workflow instantiation. are a workflow definition and technological infrastructure. The workflow definition describes the aspects for coordina- The benefits of the LODFlow include: tion, processing and controlling of the resource lifecycle. The technological infrastructure provides means for efficient de- sign and implementation of workflows. To facilitate Workflow 1. Explicitness - the declarative description of Linked Management, the technological infrastructure should fulfill Data production and maintenance workflows makes the the following requirements: activities and processes explicit and easy to understand for Linked Data engineers in a high level of abstraction. • be component-oriented, thus enabling integration and 2. Reusability - the fine-grained definition of activities interoperability among loosely-coupled system and ser- and workflow steps facilitates the reuse across work- vice components; flows. • support the implementation of a workflow, facilitating 3. Repeatability - workflows can be easily executed over resource provenance and reproducibility; and over again, thus making them repeatable, facilitat- ing testing and reliability. • ensure the correctness and reliability of applications in the presence of concurrency and failures; 4. Efficiency - the automatic workflow execution greatly • support the evolution, replacement, and addition of the reduces the required manual human intervention and workflow steps in a reengineering process. thus improves efficiency. 5. Ease of use - the high-level declarative description In a Linked Data environment, the Workflow Management of workflows and the generation of comprehensive exe- should cover the specification, execution, registration, and cution reports simplifies the definition, execution and control of procedures for producing and maintaining Linked analysis of Linked Data production and maintenance Data datasets. Therefore, in a Linked Data context, we can workflows. define the requirements for LDWFM as follows. As a result, we hope that LODFlow and LDWPO will improve Requirement 1. Linked Data Workflow Planning { the creation and maintenance of Linked Data. the capability for describing Linked Data datasets production strategy. Achieved by specifying a list of steps, where each The rest of the paper is structured as follows. We discuss step corresponds to a tool, tool configuration, as well as input Linked Data Workflow Management and its requirements and output datasets. in Section 2. In Section 3 we present the LODFlow. The evaluation of LODFlow is discussed in Section 4, based on a large-scale, real-world use case. In Section 5 we outline Requirement 2. Linked Data Workflow Execution { related work. Finally, in Section 6 we conclude the paper the capability for automating workflows, which involves a and outline directions for the future work. controlled environment for a plan execution. Achieved by running tools with tool configurations over input datasets. 2. LINKED DATA WORKFLOW MANAGE- MENT Requirement 3. Linked Data Workflow Reusability The creation and maintenance of Linked Data datasets re- { the capability to reuse the whole or a part of existing work- quire substantial efforts and resources. These efforts should flows. be planned, follow best practices, and be performed in a systematic way. The concept of Linked Data Workflow Man- Requirement 4. Linked Data Workflow Documen- agement (LDWFM) described in this section provides a for- tation { the capability to represent Linked Data Workflow mal grounding for the systematic rendering of Linked Data plans and executions in human-readable formats. management processes. The core concept of LDWFM is a Workflow. The Work- Requirement 5. Linked Data Workflow Repeatabil- flow concept originated in the business management domain, ity { the capability for reproducing the same result over time. 3. LODFLOW - THE LINKED DATA WORK- LODFlow is designed in compliance with the Linked Data FLOW MANAGEMENT SYSTEM Lifecycle [2]. Therefore, it supports the orchestration of In this section, we describe our Linked Data Workflow Man- Linked Data tools for
Load more

Recommended publications
  • Communicate the Future PROCEEDINGS HYATT REGENCY ORLANDO 20–23 MAY | ORLANDO, FL and Summit.Stc.Org #STC18
    Communicate the Future PROCEEDINGS HYATT REGENCY ORLANDO 20–23 MAY | ORLANDO, FL www.stc.org and summit.stc.org #STC18 SOCIETY FOR TECHNICAL COMMUNICATION 1 Efficiency exemplified Organizations globally use Adobe FrameMaker (2017 release) Request demo to transform the way they create and deliver content Accelerated turnaround time for 70% reduction in printing and paper customized publications material cost Faster creation and delivery of content 50% faster production of PDF for new products across devices documentation 20% faster development of Accelerated publishing across formats course content Increased efficiency and reduced 20% improvement in process translation costs while producing multilingual manuals efficiency For a personalized demo or questions, write to us at [email protected] © 2018 Adobe Systems Incorporated. All rights reserved. The papers published in these proceedings were reproduced from originals furnished by the authors. The authors, not the Society for Technical Communication (STC), are solely responsible for the opinions expressed, the integrity of the information presented, and the attribution of sources. The papers presented in this publication are the works of their respective authors. Minor copyediting changes were made to ensure consistency. STC grants permission to educators and academic libraries to distribute articles from these proceedings for classroom purposes. There is no charge to these institutions, provided they give credit to the author, the proceedings, and STC. All others must request permission. All product and company names herein are the property of their respective owners. © 2018 Society for Technical Communication 9401 Lee Highway, Suite 300 Fairfax, VA 22031 USA +1.703.522.4114 www.stc.org Design and layout by Avon J.
    [Show full text]
  • Efficient Support for Data-Intensive Scientific Workflows on Geo
    Efficient support for data-intensive scientific workflows on geo-distributed clouds Luis Eduardo Pineda Morales To cite this version: Luis Eduardo Pineda Morales. Efficient support for data-intensive scientific workflows ongeo- distributed clouds. Computation and Language [cs.CL]. INSA de Rennes, 2017. English. NNT : 2017ISAR0012. tel-01645434v2 HAL Id: tel-01645434 https://tel.archives-ouvertes.fr/tel-01645434v2 Submitted on 13 Dec 2017 HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. A Lucía, que no dejó que me rindiera. Acknowledgements One page is not enough to thank all the people who, in one way or another, have helped, taught, and motivated me along this journey. My deepest gratitude to them all. To my amazing supervisors Gabriel Antoniu and Alexandru Costan, words fall short to thank you for your guidance, support and patience during these 3+ years. Gabriel: ¡Gracias! For your trust and support went often beyond the academic. Alex: it has been my honor to be your first PhD student, I hope I met the expectations. To my family, Lucía, for standing by my side every time, everywhere; you are the driving force in my life, I love you.
    [Show full text]
  • Apache Taverna
    Apache Taverna hp://taverna.incubator.apache.org/ Donal Fellows San Soiland-Reyes @donalfellows @soilandreyes [email protected] [email protected] hp://orcid.org/0000-0002-9091-5938 hp://orcid.org/0000-0001-9842-9718 Alan R Williams @alanrw [email protected] hp://orcid.org/0000-0003-3156-2105 This work is licensed under a Creave Commons Collaboraons Workshop 2015-03-26 A,ribuIon 3.0 Unported License. Taverna Workflow Ecosystem • Workflow Language — SCUFL2 (and t2flow) • Workflow Engine — Taverna • Used in… – Taverna Command Line Tool – Taverna Server – Taverna WorkBench • Allied services – myExperiment, workflow repository – Service Catalographer, service catalog software • Instantiated as BioCatalogue, BiodiversityCatalogue, … NERSC Workflow Day 2 Map of the Taverna Ecosystem Taverna Ruby client Taverna Online Applicaon-Specific Portals Player UI UI Plugins UI Plugins Plugins Taverna UI Plugins Lite Taverna Taverna Workbench UI Plugins Command UI Plugins Line Tool REST API Components Taverna Other TavernaUI Plugins APIs Server Servers Taverna SOAP API Taverna Core AcIvity Engine UI Ps UI Plugins Plugins Workflow Service many Repository Catalogs services… 3 Users, Scientific Areas, Projects Taverna In Use NERSC Workflow Day 4 Taverna Users Worldwide NERSC Workflow Day 5 Taverna Uses — Scientific Areas • Biodiversity — BioVeL project • Digital Preservation — SCAPE project • Astronomy — AstroTaverna product • Solar Wind Physics — HELIO project • In silico Medicine — VPH-Share project NERSC Workflow Day 6 Biodiversity: BioVeL • Virtual e-LaBoratory for
    [Show full text]
  • A Review of Scalable Bioinformatics Pipelines
    Data Sci. Eng. (2017) 2:245–251 https://doi.org/10.1007/s41019-017-0047-z A Review of Scalable Bioinformatics Pipelines 1 1 Bjørn Fjukstad • Lars Ailo Bongo Received: 28 May 2017 / Revised: 29 September 2017 / Accepted: 2 October 2017 / Published online: 23 October 2017 Ó The Author(s) 2017. This article is an open access publication Abstract Scalability is increasingly important for bioin- Scalability is increasingly important for these analysis formatics analysis services, since these must handle larger services, since the cost of instruments such as next-gener- datasets, more jobs, and more users. The pipelines used to ation sequencing machines is rapidly decreasing [1]. The implement analyses must therefore scale with respect to the reduced costs have made the machines more available resources on a single compute node, the number of nodes which has caused an increase in dataset size, the number of on a cluster, and also to cost-performance. Here, we survey datasets, and hence the number of users [2]. The backend several scalable bioinformatics pipelines and compare their executing the analyses must therefore scale up (vertically) design and their use of underlying frameworks and with respect to the resources on a single compute node, infrastructures. We also discuss current trends for bioin- since the resource usage of some analyses increases with formatics pipeline development. dataset size. For example, short sequence read assemblers [3] may require TBs of memory for big datasets and tens of Keywords Pipeline Á Bioinformatics Á Scalable Á CPU cores [4]. The analysis must also scale out (horizon- Infrastructure Á Analysis services tally) to take advantage of compute clusters and clouds.
    [Show full text]
  • Programming Models to Support Data Science Workflows
    UNIVERSITAT POLITÈCNICA DE CATALUNYA (UPC) BARCELONATECH COMPUTER ARCHITECTURE DEPARTMENT (DAC) Programming models to support Data Science workflows PH.D. THESIS 2020 | SPRING SEMESTER Author: Advisors: Cristián RAMÓN-CORTÉS Dra. Rosa M. BADIA SALA VILARRODONA [email protected] [email protected] Dr. Jorge EJARQUE ARTIGAS [email protected] iii ”Apenas él le amalaba el noema, a ella se le agolpaba el clémiso y caían en hidromurias, en salvajes ambonios, en sustalos exas- perantes. Cada vez que él procuraba relamar las incopelusas, se enredaba en un grimado quejumbroso y tenía que envul- sionarse de cara al nóvalo, sintiendo cómo poco a poco las arnillas se espejunaban, se iban apeltronando, reduplimiendo, hasta quedar tendido como el trimalciato de ergomanina al que se le han dejado caer unas fílulas de cariaconcia. Y sin em- bargo era apenas el principio, porque en un momento dado ella se tordulaba los hurgalios, consintiendo en que él aprox- imara suavemente sus orfelunios. Apenas se entreplumaban, algo como un ulucordio los encrestoriaba, los extrayuxtaba y paramovía, de pronto era el clinón, la esterfurosa convulcante de las mátricas, la jadehollante embocapluvia del orgumio, los esproemios del merpasmo en una sobrehumítica agopausa. ¡Evohé! ¡Evohé! Volposados en la cresta del murelio, se sen- tían balpamar, perlinos y márulos. Temblaba el troc, se vencían las marioplumas, y todo se resolviraba en un profundo pínice, en niolamas de argutendidas gasas, en carinias casi crueles que los ordopenaban hasta el límite de las gunfias.” Julio Cortázar, Rayuela v Dedication This work would not have been possible without the effort and patience of the people around me.
    [Show full text]
  • ADMI Cloud Computing Presentation
    ECSU/IU NSF EAGER: Remote Sensing Curriculum ADMI Cloud Workshop th th Enhancement using Cloud Computing June 10 – 12 2016 Day 1 Introduction to Cloud Computing with Amazon EC2 and Apache Hadoop Prof. Judy Qiu, Saliya Ekanayake, and Andrew Younge Presented By Saliya Ekanayake 6/10/2016 1 Cloud Computing • What’s Cloud? Defining this is not worth the time Ever heard of The Blind Men and The Elephant? If you still need one, see NIST definition next slide The idea is to consume X as-a-service, where X can be Computing, storage, analytics, etc. X can come from 3 categories Infrastructure-as-a-S, Platform-as-a-Service, Software-as-a-Service Classic Cloud Computing Computing IaaS PaaS SaaS My washer Rent a washer or two or three I tell, Put my clothes in and My bleach My bleach comforter dry clean they magically appear I wash I wash shirts regular clean clean the next day 6/10/2016 2 The Three Categories • Software-as-a-Service Provides web-enabled software Ex: Google Gmail, Docs, etc • Platform-as-a-Service Provides scalable computing environments and runtimes for users to develop large computational and big data applications Ex: Hadoop MapReduce • Infrastructure-as-a-Service Provide virtualized computing and storage resources in a dynamic, on-demand fashion. Ex: Amazon Elastic Compute Cloud 6/10/2016 3 The NIST Definition of Cloud Computing? • “Cloud computing is a model for enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction.” On-demand self-service, broad network access, resource pooling, rapid elasticity, measured service, http://nvlpubs.nist.gov/nistpubs/Legacy/SP/nistspecialpublication800-145.pdf • However, formal definitions may not be very useful.
    [Show full text]
  • Workspace - a Scientific Workflow System for Enabling Research Impact
    22nd International Congress on Modelling and Simulation, Hobart, Tasmania, Australia, 3 to 8 December 2017 mssanz.org.au/modsim2017 Workspace - a Scientific Workflow System for enabling Research Impact D. Watkinsa, D. Thomasa, L. Hethertona, M. Bolgera, P.W. Clearya a CSIRO Data61, Australia Email: [email protected] Abstract: Scientific Workflow Systems (SWSs) have become an essential platform in many areas of research. Compared to writing research software from scratch, they provide a more accessible way to acquire data, customise inputs and combine algorithms in order to produce research outputs that are reproducible. Today there are a number of SWSs such as Apache Taverna, Galaxy, Kepler, KNIME, Pegasus and CSIRO’s Workspace. Depending on your definition you may also consider environments such as MATLAB or programming languages with dedicated scientific support, such as Python with its SciPy and NumPy libraries, to be SWSs. All address different subsets of requirements, but generally attempt to address at least the following four to varying degrees: 1) improving researcher productivity, 2) providing the ability to create reproducible workflows, 3) enabling collaboration between different research teams, and 4) providing some aspect of portability and interoperability - either between different sciences or computing environments, or both. Most SWSs provide a generic set of capabilities such as a graphical tool to construct, save, load, and edit workflows, and a set of well proven functions, such as file I/O, and an execution environment. Workspace is an SWS that has been under continuous development at CSIRO since 2005. Its development is guided by four core themes; Analyse, Collaborate, Commercialise and Everywhere.
    [Show full text]
  • Methods Enabling Portability of Scientific Workflows
    METHODS ENABLING PORTABILITY OF SCIENTIFIC WORKFLOWS A Dissertation Submitted to the Graduate School of the University of Notre Dame in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy by Nicholas Hazekamp Douglas Thain, Director Graduate Program in Computer Science and Engineering Notre Dame, Indiana December 2019 c Copyright by Nicholas Hazekamp 2019 All Rights Reserved METHODS ENABLING PORTABILITY OF SCIENTIFIC WORKFLOWS Abstract by Nicholas Hazekamp Scientific workflows are common and powerful tools used to elevate small scale analysis to large scale distributed computation. They provide ease of use for domain scientists by supporting the use of applications as they are, partitioning the data for concurrency instead of the application. However, many of these workflows are written in a way that couples the scientific intention with the specificity of the execution environment. This coupling limits the flexibility and portability of the workflow, requiring the workflow to be re-engineered for each new dataset or site. I propose that workflows can be written for pure scientific intent, with the id- iosyncrasies of execution resolved at runtime using workflow abstractions. These abstractions would allow workflows to be quickly transformed for different configura- tions, specifically handling new datasets, diverse sites, and different configurations. I examine three methods for developing workflow abstraction on static workflows, apply these methods to a dynamic workflow, and propose an approach that separates the user from the distributed environment. In developing these methods for static workflows I first explored Dynamic Work- flow Expansion, which allows workflows to be quickly adapted for new and diverse datasets. Then I describe an algorithm for statically determining a workflow's stor- age needs, which is used at runtime to prevent storage deadlocks.
    [Show full text]
  • Open Research Online Oro.Open.Ac.Uk
    Open Research Online The Open University’s repository of research publications and other research outputs Knowledge Components and Methods for Policy Propagation in Data Flows Thesis How to cite: Daga, Enrico (2018). Knowledge Components and Methods for Policy Propagation in Data Flows. PhD thesis The Open University. For guidance on citations see FAQs. c 2017 The Author https://creativecommons.org/licenses/by-nc-nd/4.0/ Version: Version of Record Link(s) to article on publisher’s website: http://dx.doi.org/doi:10.5281/zenodo.1308710 Copyright and Moral Rights for the articles on this site are retained by the individual authors and/or other copyright owners. For more information on Open Research Online’s data policy on reuse of materials please consult the policies page. oro.open.ac.uk Knowledge Components and Methods for Policy Propagation in Data Flows ENRICO DAGA Knowledge Media Institute Faculty of Science, Technology, Engineering and Mathematics The Open University This dissertation is submitted for the degree of Doctor of Philosophy October 2017 2 Abstract Data-oriented systems and applications are at the centre of current developments of the World Wide Web (WWW). On the Web of Data (WoD), information sources can be accessed and processed for many purposes. Users need to be aware of any licences or terms of use, which are associated with the data sources they want to use. Conversely, publishers need support in assigning the appropriate policies alongside the data they distribute. In this work, we tackle the problem of policy propagation in data flows - an expression that refers to the way data is consumed, manipulated and produced within processes.
    [Show full text]
  • HPC-ABDS High Performance Computing Enhanced Apache Big Data Stack
    HPC-ABDS High Performance Computing Enhanced Apache Big Data Stack Geoffrey C. Fox, Judy Qiu, Supun Kamburugamuve Shantenu Jha, Andre Luckow School of Informatics and Computing RADICAL Indiana University Rutgers University Bloomington, IN 47408, USA Piscataway, NJ 08854, USA fgcf, xqiu, [email protected] [email protected], [email protected] Abstract—We review the High Performance Computing En- systems as they illustrate key capabilities and often motivate hanced Apache Big Data Stack HPC-ABDS and summarize open source equivalents. the capabilities in 21 identified architecture layers. These The software is broken up into layers so that one can dis- cover Message and Data Protocols, Distributed Coordination, Security & Privacy, Monitoring, Infrastructure Management, cuss software systems in smaller groups. The layers where DevOps, Interoperability, File Systems, Cluster & Resource there is especial opportunity to integrate HPC are colored management, Data Transport, File management, NoSQL, SQL green in figure. We note that data systems that we construct (NewSQL), Extraction Tools, Object-relational mapping, In- from this software can run interoperably on virtualized or memory caching and databases, Inter-process Communication, non-virtualized environments aimed at key scientific data Batch Programming model and Runtime, Stream Processing, High-level Programming, Application Hosting and PaaS, Li- analysis problems. Most of ABDS emphasizes scalability braries and Applications, Workflow and Orchestration. We but not performance and one of our goals is to produce summarize status of these layers focusing on issues of impor- high performance environments. Here there is clear need tance for data analytics. We highlight areas where HPC and for better node performance and support of accelerators like ABDS have good opportunities for integration.
    [Show full text]
  • Arxiv:2007.04939V1 [Cs.DC] 9 Jul 2020
    This manuscript has been accepted at the Future Generation Computer Sys- tems (FGCS). DOI: 10.1016/j.future.2020.07.007 This manuscript is licensed under CC-BY-NC-ND. arXiv:2007.04939v1 [cs.DC] 9 Jul 2020 1 A Programming Model for Hybrid Workflows: combining Task-based Workflows and Dataflows all-in-one Cristian Ramon-Cortesa, Francesc Lordana, Jorge Ejarquea, Rosa M. Badiaa aBarcelona Supercomputing Center (BSC) Abstract In the past years, e-Science applications have evolved from large-scale sim- ulations executed in a single cluster to more complex workflows where these simulations are combined with High-Performance Data Analytics (HPDA). To implement these workflows, developers are currently using different patterns; mainly task-based and dataflow. However, since these patterns are usually man- aged by separated frameworks, the implementation of these applications requires to combine them; considerably increasing the effort for learning, deploying, and integrating applications in the different frameworks. This paper tries to reduce this effort by proposing a way to extend task- based management systems to support continuous input and output data to enable the combination of task-based workflows and dataflows (Hybrid Work- flows from now on) using a single programming model. Hence, developers can build complex Data Science workflows with different approaches depending on the requirements. To illustrate the capabilities of Hybrid Workflows, we have built a Distributed Stream Library and a fully functional prototype extending COMPSs, a mature, general-purpose, task-based, parallel programming model. The library can be easily integrated with existing task-based frameworks to provide support for dataflows. Also, it provides a homogeneous, generic, and simple representation of object and file streams in both Java and Python; en- abling complex workflows to handle any data type without dealing directly with the streaming back-end.
    [Show full text]
  • Intelligent Integrated Digital Platform Insilicokdd in Support of Precision Medicine for Scientific Research
    International Journal of Advanced Science and Technology Vol. 29, No. 7s, (2020), pp. 2101-2109 Intelligent Integrated Digital Platform InSilicoKDD in Support of Precision Medicine for Scientific Research [1]Plamenka Borovska, [2]Iliyan Kordev, [3]Boris Borowsky [1,2,3] Technical University of Sofia, Bulgaria Abstract The paper presents intelligent integrated digital platform InSilicoKDD based on scientific workflows in support of precision medicine designed, implemented and deployed at the Faculty of Applied Mathematics and Informatics, Technical University of Sofia. The platform is built upon our method for adaptive in silico knowledge data discovery and intelligent decision making as well as integrated approach to combine and analyze digital data at all stages of breast cancer progress, deploying modern in silico KDD methods and technologies. The conceptual architectural design of InSilicoKDD platform is presented and the software prototype revealed. Functionality and efficiency of InSilicoKDD platform have been verified on the basis of experimentation for the breast cancer issue by configuring and executing scientific workflows encompassing machine learning models for diagnostics and prediction, data sets with results of the biopsy of the cells in the patient's tumor as well as breast cancer risk assessment models. Index Terms—data intensive scientific discovery, in silico knowledge data discovery, precision medicine, scientific workflows platform. I. THE PROBLEM AREA Nowadays personalized and precision medicine [1],[2] are innovative frontier paradigms worldwide and imply adapting and optimizing of medical treatment to the individual characteristics of each patient including genetic specifics, environmental factors and risk factors of life style. Precision medicine has become possible because more people's genomes can be sequenced and more metadata collected for multiple existing therapies, but it is assumed that the most likely path to success comes from analyzing well-defined subgroups of patients.
    [Show full text]