Confirmed Tutorials

Building a Condor Desktop Grid

Michela Thiémard, Pascal Jermini, École Polytechnique Fédérale de Lausanne, Domaine IT, IT Services

Abstract

A lot of companies or schools having a large amount of desktops wish to harvest computing cycles in order to use them for research projects. Condor is a very flexible and non intrusive software available on various operating systems, that allows this cycle harvesting. It contains a lot of mechanisms in order to exploit as best as possible the available resources without disturbing or interrupting the work of the computer owner. Condor has been chosen at EPFL to build the local Desktop Grid constituted at present of about 600 cores and since 1st of Sept., 2006, 2'500'000 computing hours have been already harvested. In this tutorial we will describe the essential steps to build a Condor pool.

Outline

• Condor Project presentation
• General architecture of a Condor pool
• Condor installation on various platforms
• Condor configuration
• basic daemon parameters
• authentication and security
• compute node availability definition
• compute node software characteristics
• Submitting a job
• Script jobs, Checkpointing, Direct Acyclic Graph Manager

Presenters

Michela Thiémard obtained a Master in Mathematics in 1998 at École Polytechnique Fédérale de Lausanne (EPFL). She also obtained a Ph.D. in Applied Mathematics (Operations Research) - "On School Bus Routing and Scheduling Problem" - in 2004 at EPFL. Since 2005, she works at EPFL IT Services. She is in charge of a computing grid deployment through EPFL campus and she is the coordinator of the IT Services HPC group.

Pascal Jermini obtained his Master in Computer Science in 2006 at Ecole Polytechnique Fédérale de Lausanne (EPFL). Since 2006, he works at EPFL IT Services. His two main tasks are deploying a computing grid environment on the campus and managing the Blue Gene/L supercomputer of EPFL.

An introduction to Volunteer Computing using BOINC

Nicolas Maire, Swiss Tropical Institute, Basel, Switzerland

Abstract

Volunteer computing is an arrangement in which volunteers provide computing resources to research projects with a need for those resources. This approach was pioneered by the Folding@home and SETI@home projects, and has supplied scientific research with more computing power than any other type of computing. BOINC - the Berkeley Open Infrastructure for Network Computing - is a popular open-source platform used by dozens of research projects. BOINC addresses the challenges posed by the volunteer computing environment, for example the lack of mutual accountability between volunteers and projects, and the constraints of the commodity Internet.

This tutorial will introduce the concepts and technologies behind volunteer computing. We will discuss requirements for computationally intensive science applications to make use of this approach. The tutorial will include an introduction to BOINC, and a demonstration of setting up a BOINC-based volunteer computing project.

Outline

An overview of volunteer computing:

• Basic concepts and structure
• Application suitability and resource requirements
• Volunteer computing challenges
• Supporting communities and incentives for volunteers
• Future directions of volunteer computing

An overview of BOINC:

• BOINC concepts and features
• Client-server interaction
• Server architecture: directory structure, database, daemons and tasks, web front end
• Porting of science applications to BOINC: BOINC-API, BOINC-Wrapper, project-specific server components
• Job submission: Tools and templates
• BOINC for desktop and campus grids

Demonstration of BOINC installation and project setup:

• Prerequisites
• Project setup
• Deployment of a sample application
• Customization: Project-specific configuration, web content, and performance tuning
• Administration and monitoring
• BOINC documentation and resources

Presenter

Nicolas Maire is a member of the Biostatistics group in the Department of Public Health and Epidemiology at the Swiss Tropical Institute, where he has been working on simulation models of malaria epidemiology since 2003. He is also responsible for the development and maintenance of malariacontrol.net, a BOINC-based project that studies and refines simulation models of malaria transmission. . His background is in Biology and Computer Science. He received a PhD in Epidemiology – Stochastic simulation models of Plasmodium falciparum epidemiology and control – from the University of Basel in 2008. He has contributed to the development of the BOINC software, and is regularly involved in teaching BOINC and Volunteer Computing concepts.

Grid Metadata Management: GRelC Project and CMCC Metadata Handling System

Sandro Fiore, Euro-Mediterranean Centre for Climate Change (CMCC) and University of Salento, Lecce, Italy

Abstract

Grids encourage and promote the publication, sharing and integration of scientific data, distributed across Virtual Organizations. Scientists and researchers (climate change, bioinformatics, astrophysics, etc.) work on huge, complex and growing datasets. The complexity of metadata management in a grid environment comes from the distribution, heterogeneity and number of data sources. Along with coarse grained services (basically grid storages, replica services, storage resource managers, etc), there is a strong interest on fine grained ones concerning, for instance, grid metadata access, integration and management. As grid computing technologies and standards evolve, more mature environment (production grids such as EGEE, Teragrid, DEISA, etc.) become available for production activities and tools/services able to access in grid to relational/XML databases are also strongly required. In the proposed tutorial we will talk in detail about grid metadata management, more specifically we’ll discuss in detail about the GRelC Project.

Moreover a particular GRelC-based use case for climate change metadata management will be discussed: the Euro-Mediterranean Centre for Climate Change (CMCC) Grid Metadata Handling System (GMHS). Main topics: grid database access, management and integration; database virtualization in grid; security, performance and interoperability issues in large scale environments; real case studies concerning general purpose grid metadata management (GRelC Project) and a climate change context (CMCC Initiative).

Outline

Part I

• Basics on Database Management Systems & Grids
• Database virtualization: Grid-DB concept
• Data access and integration: existing and novel approaches
• Security issues: GSI, ACL, VO Membership management systems, etc.
• Data Grid Portals

Part II

• GRelC Project & GRelC DAS/DAIS
• Main Requirements and architecture
• Authorization: ACL and VOMS
• Grid Queries: data access and integration
• GRelC & gLite
• GRelC Portal & GILDA Tutorials

Part III

• Climate change and metadata issues
• Grid metadata management
• The Euro-Mediterranean Centre for Climate Change (CMCC) Initiative: data management issues
• CMCC Data Grid Portal
• International grid testbed on metadata management among European centres

Presenter

Sandro Luigi Fiore was born in Galatina (Italy) in 1976. He received a summa cum laude Laurea degree in Computer Engineering from the University of Lecce (Italy), in 2001. He received a PhD degree in Informatic Engineering – "Innovative Materials and Technologies" at the ISUFI-University of Lecce, Italy, in 2004. His research activity focuses on Parallel and Distributed Computing, specifically with regard to "Advanced Data Management in a Grid environment". Since 2001 he is the Project Principal Investigator of the Grid Relational Catalog project (http://grelc.unile.it) at the University of Salento, in Lecce, Italy. He was directly involved in the EGEE, EGEE-II projects ("Enabling Grids for E-science") and currently he is involved in the EGEE-III project and in other national (on bioinformatics, LIBI) and international (on grid storage management, INTERSTORE) projects. Since June 2006, he leads the Data Management Research Group at the Euro-Mediterranean Centre for Climate Change (CMCC) in Lecce (Italy). He is author and co-author of about 50 papers in refereed journals/proceedings on parallel & grid computing as well as of 1 patent concerning advanced data management. He is a member of ACM.

Trusted Computing for Trusted Infrastructures

Andrew Martin, University of Oxford, Boris Balacheff, HP Labs

Abstract

This tutorial will introduce emerging technologies designed to provide hardware-based security assurance for the implementation of infrastructure applications. Trusted infrastructure technologies have been developed based upon trusted computing, and we discuss how it may be used to enhance security, within the scope of the topics of the conference. This will include a discussion on the philosophical notions of trust as they apply to social or technological constructs and introduce a working definition for our present purposes. We will describe the main functionality of the Trusted Platform Module (TPM), the ways in which operating systems may use it to leverage new capabilities such as remote attestation, and how these may interact with capabilities for trusted networks and storage. We will describe how these ideas may in principle be applied in grid computing scenarios, and review existing research in this area.

Outline

• Security, trust, and the notion of trusted system. We will discuss notions of trust in computing, trust anchors and trust chains, and introduce some of the industry motivations to design technology for trusted systems and trust properties
• Trusted Computing and Trusted platform module (TPM). We will present the trusted computing architecture defined by the Trusted Computing Group, and discuss capabilities of the TPM; We will then explore its use in bootstrapping trusted operating systems and in particular how it can be integrated to virtualization technology to achieve so-called trusted virtualization.
• Existing applications. Trusted Network connect as an example of how network access control can be strengthened by the use of attestation; Trusted Storage as an example of how trusted platforms can substantially improve the strength of key management.
• Prototypes and Prospects. Existing R&D projects, providing substantial demonstrators of tools for trusted infrastructure; architectural ideas for trusted grid computing, and the prospects for its implementation. We will use specific examples and present a demonstrator from the current Open Trusted Computing European project (www.opentc.org).

Presenters

Andrew Martin has been lecturing for over eight years in the Computer Science Faculty at the University of Oxford and there leads a Master's programme in Software and Systems Security. He teaches a full module on Trusted Computing within that programme, and has also lectured at European and Asian Summer Schools in this area. He has participated in a number of Oxford's Grid projects, supervises several doctoral students researching ways to apply Trusted Computing in Grid and Service-oriented contexts, and was formerly a Research Group chair within the Global (now Open) Grid Forum.

Boris Balacheff is a HP Labs expert in the field of computer security, specializing in the area of trusted computing and trusted infrastructure technologies. He sits on the Board of Directors of the Trusted Computing Group (TCG) and co-chairs its Certification Program Committee. Boris also serves on HP’s corporate Security Office where he focuses on HP’s trusted infrastructure security strategy. He is one of the early contributors to the invention of Trusted Computing technology, and he co-authored the HPLabs’ book on this topic. He also served on the Technical Committee of the Trusted Computing Platform Alliance (TCPA) during the development of its early specifications. He has helped to facilitate a number of Summer Schools on this material, presenting at some of them, and regularly presents these ideas to HP's customers.