1. SC2004 GLIF Panel Peter Clarke GLIF Infrastructure SC2004 Panel Discussion Peter Clarke UK National e-Science Centre

2. SC2004 GLIF Panel Peter Clarke Questions Why are Canada, Japan, the Netherlands, the United Kingdom and the United States involved in GLIF? What can be accomplished on lambdas that cannot be accomplished on "best-effort" networks? Where does your infrastructure connect? What's missing? When will we implement a totally functional LambdaGrid? What do you see as the future of networking?

3. SC2004 GLIF Panel Peter Clarke Questions Why are Canada, Japan, the Netherlands, the United Kingdom and the United States involved in GLIF? Community Lobby National Competitiveness in Research Acceptance that one size manifestly doesnt fit all Capability for Research You cant do it alone (by definition its an end-2-end thing) Trust

4. SC2004 GLIF Panel Peter Clarke Initial condition: Random water/ surfactant mixture. Self-assembly starts. Rewind and restart from checkpoint. Lamellar phase: surfactant bilayers between water layers. Cubic micellar phase, low surfactant density gradient. Capability : Steering & Migration of Computationally intensive simulations

5. SC2004 GLIF Panel Peter Clarke UK HPC machines connected to TeraGrid for SC2003 Now a persistent connection Source: Reality Grid Project www.realitygrid.org

6. SC2004 GLIF Panel Peter Clarke Capability: Breast Screening Programmes Current system is minimal technologically (taken from e-DiaMoND Project)

7. SC2004 GLIF Panel Peter Clarke NETWORK Future should be digitised and processed images Remote Radiographers ? Remote Patient information 2 nd opinion Requires ~ Gbit/s flows for remote access Will not be possible without scheduled guaranteed net-services (taken from e-DiaMoND Project)

8. SC2004 GLIF Panel Peter Clarke ~1-10 Gbps Tier2 Center Online System CERN USA Centres France Center ITALY Center Institut e ~20 PBytes/year ~PByte/sec ~1-10 Gbps Tier2 Center Experiment CCLRC-RAL Centre GEANT Network SuperJANET ~1-10 Gbps

9. SC2004 GLIF Panel Peter Clarke What can be accomplished on lambdas that cannot be accomplished on "best-effort" networks? An R&D network is not a production network – you can do out of the box things Certainty of having what you need when you need it (but not scalable today of course) – means research computing can be planned Ability to use non standard transport Fixed latency => backplane connections between compute centres => virtual machine rooms Provide virtual networks a la UCLP, and push management to the users

10. SC2004 GLIF Panel Peter Clarke CDF – Connection UCL-FNAL

11. SC2004 GLIF Panel Peter Clarke Where does your infrastructure connect ?

12. SC2004 GLIF Panel Peter Clarke 10 Gbit/s SURFnet 10 Gbit/s SURFnet 10 Gbit/s IEEAF 10 Gbit/s Prague CzechLight Prague CzechLight 2.5 Gbit/s NSF 10 Gbit/s Stockholm NorthernLight Stockholm NorthernLight CA*net4 2.5 Gbit/s New York MANLAN New York MANLAN Tokyo WIDE Tokyo WIDE 10 Gbit/s IEEAF 10 Gbit/s 2.5 Gbit/s Tokyo APAN Tokyo APAN Amsterdam NetherLight Amsterdam NetherLight Geneva CERN Geneva CERN London UKLight London UKLight Chicago StarLight Chicago StarLight Source: Kees Neggers, SURFnet

13. SC2004 GLIF Panel Peter Clarke As a result of the initiative, UKERNA have pushed through a complementary initiative to provide a nationwide domestic R&D network in parallel with SupetJANET LeNSESWERN Kentish MAN NorMAN YHMAN EMMAN EastNet LMN South Wales MAN TVN MidMAN Northern Ireland NNW C&NL MAN GlasgowEdinburgh WarringtonLeeds Reading London Bristol Portsmouth EaStMAN UHI Network Clydenet AbMAN FaTMAN S T C T T T TT T S S S T T London UKLight London UKLight

14. SC2004 GLIF Panel Peter Clarke What's missing? When will we implement a totally functional LambdaGrid? When NRENS all adopt hybrid networking as the normal business model o Requirement: irrefutable proof of benefit o Political: national competitiveness, one price for all issues.. o Financial: proof of cost effectiveness o Policy: When we have a common and standards based control plane infrastructure – not individual solutions : now is the time to get together on this one (IMHO) I have no idea about the last mile issue

15. SC2004 GLIF Panel Peter Clarke What do you see as the future of networking Wrong guy to ask – I think from point of view of future of computationally intensive research where I would say: We must get to point where network is not seen by researchers as a big barrier which keeps sites disjoint Must ensure it is understood that research capability is currently inhibited by traditional network provision Must ensure it is understood that one size doesnt fit all is a good idea (there is no way research could be carried out if we all took commercial IP connections)

16. SC2004 GLIF Panel Peter Clarke … a possible Network of the Future…. Provides Normal best effort IP network where appropriate Extended virtual LANs where appropriate Switched lightpaths where appropriate layer-2 split out of wavelength [e.g. Ethernet] Router Routed IP Network