1. 1 An update on the Open Science Grid for IHEPCCC Ruth Pordes, Fermilab

2. 2 OSG -- a reminder... The OSG Distributed Facility is a US grid computing infrastructure that supports scientific computing via an open collaboration of science researchers, software developers and computing, storage and network providers. OSG provides access to existing computing and storage resources contributed by members of the OSG Consortium. The OSG Consortium policies are to be open to participation by all researchers. The OSG Project is co-funded by DOE and NSF for 5 years at $6M/year starting in Sept ‘06 currently including deliverables for US LHC, LIGO and STAR; use by CDF and D0; and with potential other experiment deliverables in the future. The OSG Project responsibilities are to operate, protect, extend and support the Distributed Facility for the Consortium.

3. 3 OSG is part of the WLCG OSG is relied on by the US LHC as their Distributed Facility in the US. Resources accessible through the OSG infrastructure deliver accountable cycles for the US LHC experiments. OSG interoperates with many other infrastructures in managerial, operational and technical activities. OSG cooperates specifically with the EGEE to ensure an effective and transparent distributed system for the experiments. OSG supplies the Virtual Data Toolkit to OSG, EGEE and WLCG - a packaged, integrated and distributable set of middleware including Condor, Globus, Myproxy and components needed by the scientific community.

4. 4 What software is in the VDT? Security  VOMS (VO membership)  GUMS (local authorization)  mkgridmap (local authorization)  MyProxy (proxy management)  GSI SSH  CA CRL updater Monitoring  MonaLISA Accounting  OSG Gratia Support  Apache  Tomcat  MySQL (with MyODBC)  Non-standard Perl modules  Wget  Squid  Logrotate  Configuration Scripts Job Management  Condor (including Condor-G & Condor-C)  Globus GRAM Data Management  GridFTP (data transfer)  RLS (replication location)  DRM (storage management)  Globus RFT Information Services  Globus MDS  GLUE schema & providers  gLite CEMon Client tools  Virtual Data System  SRM clients (V1 and V2)  UberFTP (GridFTP client) Developer Tools  PyGlobus  PyGridWare Testing  NMI Build & Test  VDT Tests

5. 5 Current OSG deployment 96 Resources across production & integration infrastructures 27 Virtual Organizations including operations and monitoring groups >15,000 CPUs ~6 PB MSS ~4 PB disk

6. 6 OSG core competancies Integration: Software, Systems, Virtual Organizations. Operations: Common Support & Grid Services. Inter-Operation: Bridging Administrative & Technical Boundaries. with Validation, Verification and Diagnosis at each step. with Integrated Security Operations and Management.

7. 7 OSG support for non-physics communities Frank told you about our non-physics community activities last time. Since then: Alan Blatecky’s group at RENCI is porting the “award winning Bioportal” to OSG. >100 Nanotechnology jobs -- that run from 10-40 days -- are being executed on LIGO, ATLAS and CMS sites. We are discussing partnership with the Northwest Indiana Computing Grid (NWICG) -- which brings me to Gaussian. When we start talking to Computational Chemistry we quickly run into licencing issues. Yes, we also say it is the responsibility of the project/VO.. But there are 50 sites on OSG. The P-Grade portal has been interfaced to a version of CHARMM molecular dynamics simulation package. Some versions of this also have licencing issues. Work on Campus Grids enabling Crimson Grid, NWICG, New York State Grid (NYSG), GPN (Nebraska Education/Training grid) partnerships. (Note: Partners do not have to contribute resources; collaboration can equally be in software, procedures, education, training, security etc.)

8. 8 OSG is in a flat world OSG is one of many grids. VOs interface to more than one Grid. Computing and Storage Resources are accessible to more than one Grid. Any Work might be done using multiple grids e.g. Workflow submitted using the CMS analysis Grid interface; Dispatched using the EGEE Resource Broker; Data is transferred from an OSG site; Job scheduled through the FermiGrid Campus Grid gateway; and Executed on the local CDF grid site.

9. 9 OSG Interoperation - Security OSG Security is based on managerial, operational and technical controls that manage risk. OSG regards Site and VO Security Responsibilities as equivalent. VOs cross Grid boundaries -- so coordination is essential. OSG - EGEE Security groups are Joint. This does not mean all policies and documents are the same -- but that we work together to be consistent and in common wherever sensible. We are also working together on security middleware extensions (e.g. to allow “pull” architecture for job scheduling)

10. 10 OSG Interoperation - Software From now on software release process includes tests for interoperability. VO end-to-end systems are part of the OSG “concern”: VOs make their priorities clear for the common middleware. VOs contribute to the testing of new OSG releases. VOs consider commonality of and are prepared to contribute middleware that they develop or adopt. VOs often use (and therefore harden) new components in the VO Environment before they are part of the common middleware. TeraGrid CTSS and OSG VDT software stacks being aligned e.g. have the same Globus patches applied.

11. 11 OSG Interoperation - Jobs OSG publishes information to the WLCG information service through a web interface. Laurence combines the information with that from the EGEE to publish to the WLCG. In practice we use the same infrastructure as EGEE, with our own information gatherers. E.g. how CMS dispatches job (courtesy O. Gutsche)

12. 12 OSG Interoperation - Data OSG middleware supports the GridFTP and the SRM storage interfaces. Catalogs, Replication etc is in the scope of the VOs. There are 3 implementations of SRM on OSG: srm/dCache, Jasmine/JLAB, srm/drm (LBNL). A 4th srm/LStore is in test. Storage management is a major focus of the next year.

13. 13 OSG Interoperation - Operations Automated ticket exchange in distributed support centers in OSG. Working on automated ticket exchange with EGEE. Manual procedures of course work first! OSG EGEE

14. 14 OSG Interoperation - Education We will continue the successful iVDGL grid summer workshop (hands on training for a week). We will work more closely with the UK eScience program on the International Summer School on Grid Computing. (ISSGC) OSG works with TeraGrid on education and training activities.

15. 15 Interoperation Concerns How do we communicate and test interoperability of changes (interfaces and capabilities) before they get to production? How do we communicate about new s/w developments in time to have common approaches & avoid duplication & divergence? How do we manage ourselves to plan ahead and not do “just in time” developments. And for OSG how do we prioritize support for our non-WLCG stakeholders during data taking?

16. 16 GIN - Grid Interoperability Now OSG is part of the 9 grid partnership for Grid Interoperability Now (GIN). Tests evolved since “Worldgrid” between DataTag/EDG and Grid3 a few years ago. OSG specific contributions:  Storage interface tester across 6 implementations.  Sites for executing jobs from Application tests (e.g. Ninf-G, charmm).  Bilateral interoperation with EGEE for Information Services. Looking to work on Service Discovery with EGEE and NGDF next. http://forge.ggf.org/sf/wiki/do/viewPage/projects.gin/wiki/HomePage

17. 17 Summary The Open Science Grid has lots of work it is doing and lots more work to do!