IGI – the Italian Grid initiative and its impact for the Astrophysics community

1. IGI – the Italian Grid initiative and its impact for the Astrophysics community Fabio Pasian INAF – Information SystemsUnit INAF – Osservatorio Astronomico Trieste

2. Outline • IGI – definition and scope • Previousinitiatives • Activities • Impact on astrophysics F.Pasian - Italian Grid Infrastructure

3. IGI • IGI = association for the Italian Grid Infrastructure • Has been established as a consortium of 10 different national institutions and 4 consortia to provide long term sustainability for the Italian Grid. • Partners: INFN, CNR, ELETTRA, ENEA, INAF, INGV, UNICAL, UNINA, UNIPG, UNIPNM; COMETA, COSMOLAB, GARR, SPACI F.Pasian - Italian Grid Infrastructure

4. IGI – scope • IGI focuses on setting up and operating a common e-Infrastructure for the Italian Sciences, and includes the main public Resource Providers and Computing Centers, in addition to various Regional Initiatives and other related projects. IGI will provide a consistent and coordinated Italian strategy as a step towards the European Grid Initiative (EGI) and as an interface to the EU Grid infrastructure projects e-IRG (e-Infrastructure Reflection Group) and ESFRI (European Strategy Forum on Research Infrastructures), and to other international activities as the need arises. F.Pasian - Italian Grid Infrastructure

5. INFNGrid • The INFN Grid project, approved in late 1999, has developed and deployed the first Italian Grid Infrastructure, based on GARR, the Italian research network. • More than 30 sites, both INFN and some among the most important Italian universities. • Although primarily focused on physics, it has been, since the beginning, open to other fields of research (bio-medicine, earth observation, etc.) and to industry. • Training/support activities F.Pasian - Italian Grid Infrastructure

6. Grid.it • With a grant received from the MIUR-FIRB funds (governmental funds for investment in fundamental research) for the Grid.it project, a number of National Research Institutions paved the way in 2002-06 for the development of a common production grid infrastructure supporting the Italian Research Area. • Developments • Production Grid F.Pasian - Italian Grid Infrastructure

7. Transition • Grid.it : 50 sites ("resourcecenters"), 4500 CPUs (4.5 MSPECINT), 1 PBytestorage, 2.5 Gbpsbandwidth • Grid.it has come to a close in 2006 as a development project; to extend its benefits, IGI has taken over and acts as the national coordinator for the different pieces of the Italian e-Infrastructure present in EGEE. F.Pasian - Italian Grid Infrastructure

8. Grid.it SPACI GRISù (Cometa, Cresco, Cosmolab, Scope) Grid@Trieste F.Pasian - Italian Grid Infrastructure

9. PON consortia • National fundingtoenhancecomputationalinfrastructures in southern Italy • Fourregionalprojectscreated: • COMETA consortium (TriGrid and PI2S2 projects) → U.Becciani’s talk • Cosmolab (CyberSar project) • SCOPE (UniversityofNaples) • CRESCO (ENEA) • Share middleware (INFNGriddistribution) → interoperability • MiURresourcestoexpand PON experiencetorestof Italy F.Pasian - Italian Grid Infrastructure

10. IGI – builduponprevious work • Leverage upon the large integration and standardization effort of international and national research projects such as EDG, LCG, Grid.it and EGEE, in order to establish a coherent platform of interoperable Grid Services customized for Italian user applications, and make it available in support of the Italian scientific end user communities and, possibly, early commercial adopters F.Pasian - Italian Grid Infrastructure

11. IGI middlewaredistribution (I) • The idea istoexpand the EGEE gLitedistributionto tackle needsofItalian Grid community • TakingfromINFNGridexperience: withrespectto the CERN gLitedistribution, totallycompatible; hasadditionalcomponents and specificconfigurations. Among the specificities: • Supportfor 20 VOrg’s • StoRM(accesstostorage) • P’n’P User Interface, • Configured MPI support • GridICEmonitoring on WorkerNodes • … F.Pasian - Italian Grid Infrastructure

12. IGI middlewaredistribution (II) • Whichcomponents can go in the release? • A Board (the Executive Boardof IGI) recieves the requests and evaluates • Functionality and addedvalueofcomponent • ReferenceUser/VOrg community • Usageexperience and testsperformed • Evolution/development/sustainabilityofcomponent and decideswhich are the additionalcomponentswhich can beinserted in the release, definingpriorities • A Release Team perfoms a technicalevaluation and defines the activityplan and schedule • Tight collaborationwithdevelopersismandatory F.Pasian - Italian Grid Infrastructure

13. IGI middlewaredistribution (III) F.Pasian - Italian Grid Infrastructure

14. National framework • IGI hassubmitted a proposaltobeincludedwithin the roadmapforItalianresearchinfrastructures • Appropriate fundingtobeprovidedbyMiUR • Complementaryto HPC proposal • Coordinatedwith (part of) ESFRI roadmapfore-infrastructures F.Pasian - Italian Grid Infrastructure

15. Communities • IGI plans to support activities in a vast range of scientific disciplines — e.g. Physics, Astrophysics, Biology, Health, Chemistry, Geophysics, Economy, Finance — and any possible extensions to other sectors such as Civil Protection, e-Learning, dissemination in Universities and secondary schools. F.Pasian - Italian Grid Infrastructure

16. Astrophysicsasan IGI community • Among these, the Astrophysics community is active as a user, by porting applications of various kinds, but also as a resource provider in terms of computing power and storage, and as middleware developer. F.Pasian - Italian Grid Infrastructure

17. Grid(s) forastrophysics A stepbackwards… • Dedicated WP in Grid.it • DRACO: data gridforastrophysics and coordinationwith the VObs (2003-05) • Separate projectsafterwards: VObs.it and DRACO2 (relying on the Italiangridinfrastructure) fundedby INAF F.Pasian - Italian Grid Infrastructure

18. Porting ofapplications • within the Grid.it and DRACO projects: • Montecarlo • Hi-resN-bodysimulations • Data processing (OmegaCam) • within the EGEE/EGAAP framework: • simulationsof the Planck mission • on COMETA: • amongothers, HPC applications (e.g. FLY) • on IGI (and EGEE): • populationof the BaSTIarchive (see poster) • Leadership of A&A EGEE cluster. F.Pasian - Italian Grid Infrastructure

19. Developmentofmiddleware • definitionofrequirements • developmentofgridftp+EGEEdriversfor CFITSIO • developmentof G-DSE → C.Vuerli’s talk • originally in collaborationwith INFN • “native” Grid accesstorelationaldatabases • applied (and optimised) forastrophysics • benchmarking and comparisonamongsystemsmanagingDBs on the Grid • Leadership of EGEE Database Group. F.Pasian - Italian Grid Infrastructure

20. Database accessbenchmarks F.Pasian - Italian Grid Infrastructure

21. F.Pasian - Italian Grid Infrastructure

22. from K.Baxevanidis F.Pasian - (Super)computing and e-Infrastructures

23. Appl Appl ... Appl Appl domain-specific Data Infrastructure generic Computing Infrastructure Network Infrastructure e-Infrastructure: components When defining the roadmap for the development of an e-Infrastructure for Astronomy, these components need to be thought as integrated, or at least fully interoperable F.Pasian - (Super)computing and e-Infrastructures

24. Knowledge Infrastructure VObs appl appl ... appl appl Data Infr. Computing Infr. Network Infr. This is the area where the VObs expects to bring in its own standards, tools and facilities The path forward for e-infrastructure • All facilities should be integrated in an interoperable e-infrastructure • Goals for the next years: • Consolidating a Data Infrastructure on top of Network and Computing • Domain-specific applications on top of the Data Infrastructure • Inter-operation of applications (data processing, analysis, mining, bibliography, semantic web) to build a Knowledge Infrastructure F.Pasian - Italian Grid Infrastructure