1. Virtual Laboratory for e-Science (VL-e) Henri Bal Department of Computer Science Vrije Universiteit Amsterdam bal@cs.vu.nl vrije Universiteit

2. Outline e-Science and virtual laboratories The VL-e project VL-e and networking Case studies: oVisualization oInteractive problem solving environments oDistributed supercomputing Computing/networking infrastructure

3. e-Science Web is about exchanging information Grid is about sharing resources oComputers, data bases, instruments, services e-Science supports experimental science by providing a virtual laboratory on top of Grids

4. Management of comm. & computing Management of comm. & computing Management of comm. & computing Potential Generic part Potential Generic part Potential Generic part Application Specific Part Application Specific Part Application Specific Part Virtual Laboratory Application oriented services Grid Harness multi-domain distributed resources Virtual Laboratories Distributed computing Visualization & collaboration Knowledge Data & information

5. Optical Networking High-performance distributed computing Security & Generic AAA Virtual lab. & System integration Interactive PSE Collaborative information Management Adaptive information disclosure User Interfaces & Virtual reality based visualization Bio-diversity Bio-Informatics Telescience Data Intensive Science Food Informatics Medical diagnosis & imaging Virtual Laboratory for e-Science

6. The VL-e project 40 M€ (20 M€ BSIK funding) 2004 - 2008 vrije Universiteit 20 partners Academic - Industrial

7. VL-e and networking e-Science applications generate much (distributed) data oHigh-resolution imaging oBio-informatics queries oParticle physics: oCurrently: 1 PByte per year oLHC (2007): 10-30 PByte per year Virtual laboratories need high-speed networks for oRemote visualization oInteractive problem solving environments oDistributed supercomputing

8. VL-e and networking Optical Networking High-performance distributed computing Security Virtual labi PSE CIM A.I.D. Visualization Bio-div Bio-Inf Telesc ience Data Intensie Food Medical imaging

9. Visualization on the Grid

14. MRI, PETMonolith, ClusterCave, Wall, PC, PDA From Medical Image Acquisition to Interactive Virtual Visualization… MD login and Grid Proxy creation Bypass creation LB mesh generation Job submission Job monitoring Virtual Node navigation Simulated Blood Flow Patient at MRI scanner MR image Segmentation Shear stress, velocities Simulated blood flow se (e.g., Leiden)ce (e.g., Valencia)ce (e.g., Bratislava) ui (VRE) P.M.A. Sloot, A.G. Hoekstra, R.G. Belleman, A. Tirado-Ramos, E.V. Zudilova, D.P. Shamonin, R.M. Shulakov, A.M. Artoli, L. Abrahamyan Interactive Problem Solving Environments

15. Distributed supercomputing ( parallel computing on grids) VU (72 nodes) UvA (32) Leiden (32) Delft (32) GigaPort Utrecht (32) DAS-2 Distributed ASCI Supercomputer 2

16. Distributed supercomputing ( parallel computing on grids)

17. Can grids be used for High-Performance Computing applications that are not trivially parallel? Key: grids usually are hierarchical oCollections of clusters, supercomputers oFast local links, slow wide-area links Can optimize algorithms to exploit this hierarchy oMessage combining + latency hiding on wide-area links oOptimized collective communication operations (broadcast etc.) oOften gives latency-insensitive, throughput-bound algorithms HPC on a grid?

18. Ibis: a Java-centric grid programming environment Written in pure Java, runs on heterogeneous grids o“Write once, run everywhere ” Many applications: oElectromagnetic simulation (Jem3D) oAutomated protein identification (VL-e application from AMOLF) oN-body simulations oSAT-solver oRaytracer Jem3D (see SC’04) Available from www.cs.vu.nl/ibis

19. Networking demands Low latency is needed for oInteractive visualization oInteractive Problem Solving Environments oSynchronous, latency-sensitive parallel algorithms High throughput is needed for oData-intensive e-Science applications oVisualization of large data sets oAsynchronous, throughput-bound parallel algorithms Efficient collective (group) communication for oCollaborative visualization between multiple sites oCollective operations in parallel algorithms

20. Outline e-Science and virtual laboratories The VL-e project VL-e and networking Examples: oVisualization oInteractive Problem Solving Environments oDistributed supercomputing Computing/networking infrastructure

21. Grid Middleware Gigaport Network Service (lambda networking) Application specific service Application Potential Generic service & Virtual Lab. services Grid & Network Services Virtual Laboratory VL-E Experimental Environment VL-E Proof of concept Environment Telescience Medical Application Bio ASP Virtual Lab. rapid prototyping (interactive simulation) Additional Grid Services (OGSA services) VL-e environments

22. DAS-3 Proposed next generation grid in the Netherlands Partners: oASCI research school (VU, UvA, TU Delft, Leiden) oGigaport-NG/SURFnet: DWDM computer backplane (dedicated optical group of 8 lambdas) oVL-e and MultimediaN BSIK projects Topology controlled by applications through the Network Operations Center

23. DAS-3DAS-3 CPU’s R R R R R NOC

24. Summary VL-e (Virtual Laboratory for e-Science) studies entire e-Science chain, including applications, middleware and grids High networking demands from applications and generic methods New state-of-the-art Grid infrastructure planned for 2006 using optical networking