15 EGEE Project

The EU 6th Framework Program project EGEE (Enabling Grids for E-sciencE) is one of the largest international projects funded by the European Union, both in the number of involved partners and in the EU financial contribution. The two-year project led by CERN started on April 1, 2004 and includes 70 partners not only from almost all European countries but also from Russia and USA, although the U. S. institutions do not receive direct financial contribution from the EU. Nowadays, some Asian partners from Korea and Japan also expressed interest in collaborating with the EGEE project.

The EU contribution of some 35 million Euro is used to build a pan European Grid that will be prepared to interoperate with similar non-European Grids. The EGEE Grid is based on a network of interconnected data warehouses and computers, mostly clusters with Intel IA-32 and IA-64 architecture based processors or compatible architectures made by AMD. The project budget does not include direct investment; on the contrary, the equipment should be contributed by interconnected countries and project partners. The project contribution is the middleware, i.e., the programs used to interconnect individual computers and data warehouses to hide the complexity of the underlying fabric and to present the Grid environment in a form as unified as possible. This middleware will be used to connect individual computers and already existing thematic, regional or national Grids into one unified pan European Grid infrastructure.

To simplify the management of such a large project, whole Europe is divided into several so-called Federations based on the regional affiliation. The Czech republic is a member of the Central European Federation whose other members are Austria, Hungary, Poland, Slovakia and Slovenia. CESNET is the only institution from the Czech Republic directly participating in the EGEE project but in fact, CESNET coordinates researchers from several other institutions, most notably the Masaryk University in Brno, Institute of Physics of the Academy of Sciences, West Bohemia University in Pilsen, and Charles University in Prague. The extent of CESNET involvement, as well as the financial contribution, makes CESNET the largest EGEE partner from the whole Central European Federation.

The regional division of partners is complemented by the division of the technical content of the project. There are four groups of activities:

Joint Research Activities (JRA)
Networking Activities (NA)
Specific Service Activities (SA)
Networking (Coordination) Activities (NA)

The Joint research activities cover the following areas:

JRA1 - middleware development, re-engineering and integration
JRA2 - quality assurance
JRA3 - security
JRA4 - advanced network services

CESNET is involved in the middleware development and re-engineering activities of the JRA1. This is the largest research activity of the EGEE project and its results will deeply influence the success of the whole project. The JRA1 work builds on the results of the preceding DataGrid (where CESNET also participated) EU 5FP project; its goal is to create a uniform environment for resource management and scheduling and for the access to the data storage and to data stored there. The complete middleware will be available under the brand name gLite.

The JRA1 activities are led by CERN with participation from institutions from the Great Britain, France, and Italy. The Czech Republic is the only other country also involved in these research activities - this success is the result of previous CESNET work within the EU DataGrid project. CESNET is responsible for the development of the Logging and Bookkeeping service. This service is responsible for monitoring the job flow through the individual middleware and other Grid fabric components. It collects events triggered by jobs passing through these components and reconstructs the state of jobs (submitted, queued, waiting, running, etc.) from these events. The Logging and Bookkeeping service was designed as a specialized Grid monitoring service, with proprietary interfaces for event logging and also for access to the database that stores both the events and job states. It is being re-engineered to provide standard grid service interfaces which will allow its more extensive use as part of a general Grid monitoring infrastructure.

The CESNET research team also designed and is currently developing the Job Provenance - a long-term storage of job related information. While the Logging and Bookkeeping service keeps track of active jobs only (the information is purged from the bookkeeping database when the job is finished or aborted), the Job Provenance will keep the information as long as necessary. The information stored in the Job Provenance could be used for statistical purposes (we collaborate with the JRA2 activity which uses the bookkeeping information to evaluate the EGEE Grid use efficiency) but also to re-run jobs (e.g., when a new algorithm is developed and the same data must be re-examined). While the Job Provenance is a new, EGEE related concept, it has been included already as a part of the first version of the gLite middleware (it is a part of the so-called Release Candidate 1).

As already mentioned above, CESNET collaborates with the JRA2 on providing input information for Grid metrics evaluation. While not directly involved in the JRA3 activities, a member of the CESNET team serves as an official liaison between the JRA1 and JRA3 groups. CESNET is not directly involved in the JRA4 activities, however, participating in the GÉANT2 project, the CESNET also contributes to the networking related work of the EGEE project.

A stable and reliable production Grid is the main goal of the EGEE project. This Grid should interconnect over 100 sites with more than 50 thousand processors and at least 1 PB of disk capacity, all available in the year 2006. It is the responsibility of the SA1 - the Grid management and operation activity - to deploy the middleware and manage the resulting Grid. Almost all EGEE partners are to some extent involved in this activity, and CESNET is no exception. We are a part of the Central European Regional Operating Centre (ROC); CESNET provides a backup for most of the services running in the ROC headquarters in Poland. The Institute of Physics (Academy of Sciences) closely cooperates with CESNET on the ROC management in the Czech Republic, and nowadays some 300 processors represent the Czech Republic contribution to the EGEE project - this is almost 50 % of all the current Central European capacity. In November 2004, CESNET took a leading role to develop and deploy a Central European Virtual Organization (VOCE) which will provide production access to the CE resources to end users not (yet) participating in any of the application oriented virtual organizations (like High energy physics, computational chemistry, bioinformatics, etc.) already established within the EGEE.

The SA2 activity - deployment of specific network services - is responsible for interconnecting the EGEE project with the GÉANT2 EU project. As with JRA4, CESNET is involved indirectly, through its participation in the GÉANT2 project.

CESNET is also participating in two of the five networking (coordination) activities: the NA3 (Training) and NA4 (Application Identification and Support). CESNET has organized two user training events already - the introductory seminar in late October and an end user training workshop in early December. We also intensively support high energy physics applications, especially through participation in the so-called Experiment Data Challenges where hundred thousands jobs are submitted and run on the already available Grid infrastructure both to get some scientific results and also to stress test the infrastructure itself. The combined CESNET and Institute of Physics contribution has been usually the largest from the whole Central European region. We also collaborate with other application groups, helping them to move their applications into the EGEE Grid environment. The most active collaboration exists with teams involved in computational chemistry.

The remaining three networking activities are the Project management (NA1), Dissemination (NA2), and International Cooperation (NA5). CESNET is not directly involved in any of these activities, although a CESNET representative is a member of Collaboration Board (a body consisting of one representative from each EGEE partner).

More information about the project can be found at its web pages. Specific information about the CESNET participation with the EGEE project can be found at the following URL: egee.cesnet.cz.

contents