automated gole pilot project update
play

Automated GOLE Pilot Project update Jerry Sobieski Oct 2012 - PowerPoint PPT Presentation

NORDUnet Nordic infrastructure for Research & Education Automated GOLE Pilot Project update Jerry Sobieski Oct 2012 Chicago, US Automated GOLE Fabric NORDUnet Nordic infrastructure for Research & Education Nordunet KRLight


  1. NORDUnet Nordic infrastructure for Research & Education Automated GOLE Pilot Project update Jerry Sobieski Oct 2012 Chicago, US

  2. Automated GOLE Fabric NORDUnet Nordic infrastructure for Research & Education Nordunet KRLight GLORIAD PSNC NetherLight CANARIE GEANT StarLight CzechLight JGN-X ACE KDDI Labs UvA Cern AIST • The GLIF Automated GOLE global fabric – The GLIF Automated GOLE Pilot was initiated in 2010 to provide a global fabric of Open Lightpath Exchanges for the express purpose of maturing the dynamic provisioning software, demonstrating the value of GOLEs to emerging network service models, and to develop a set of BCP for these services.

  3. NORDUnet Nordic infrastructure for Research & Education Ethernet Transport Service KRLight.ets GLORIAD.ets NorthernLight.ets CzechLight.ets Daejeon Chicago Copenhagen Prague DynamicKL OpenNSA OpenNSA OpenDRAC A A A KRLight KDDI-Labs.ets CESNET NORDUnet + GLORIAD Fujimino SURFnet Pionier.ets G-LAMBDA-K ACE JGN-X Pionier Poznan A A AutoBAHN JGNX.ets StarLight.ets NetherLight.ets Tokyo Chicago Amsterdam GEANT G-LAMBDA-K OpenNSA/Argia OpenDRAC A A A UvALight.ets GEANT.ets ESnet.ets AIST.ets University of Ams. Paris Chicago Tsukuba OpenNSA AutoBAHN G-LAMBDA-A OSCARS NSI Networks (“A”=Aggregator) NSI peerings (SDPs) unless otherwise indicated these are vlans 1780-1783 In-progress peerings (SDPs) or GOLEs

  4. NORDUnet Nordic infrastructure for Research & Education KRLight A Daejeon WIX GLORIAD NorthernLight.ets CzechLight.ets DynamicKL Washington Chicago Copenhagen Prague KRLight OpenNSA OpenNSA OpenNSA DRAC MAX A A A A HKOEP Hong Kong CESNET NORDUnet + KDDI-Labs SURFnet Fujimino GLORIAD CANARIE Pionier.ets G-LAMBDA-K JGN-X Poznan ACE A A AutoBAHN StarLight JGN-X Pionier NetherLight Tokyo Chicago Amsterdam OpenNSA/Argia G-LAMBDA-K DRAC GEANT US LHCnet A A UvALight.ets Esnet AIST.ets GEANT.ets University of Ams. Chicago Tsukuba Paris OpenNSA G-LAMBDA-A OSCARS AutoBAHN NSI Networks (“A”=Aggregator) NSI peerings (SDPs) unless otherwise indicated these are vlans 1780-1783

  5. EVOLUTION: NORDUnet AUTOMATED GOLE + NSI 2013 Nordic infrastructure for Research & Education NorthernLight KRLight.ets DAE A A CHI CPH A KRLight GLORIAD A NYC HAM A CSTnet CzechLight SEA CHI HKOEP BEJ PRA ASH A HKG A A A A A MANLAN PWAVE.ets CESNET A WIX KDDI-Labs.ets Pionier WDC JGN-X Fujimino POX A A A JGNX.ets StarLight ACE Tokyo Chicago NetherLight UvALight US LHCnet AMS AMS GEANT A A GEANT A AIST.ets ESnet.ets RNP.ets LON A Tsukuba NRENs A A NSI Networks (“A”=Aggregator) NSI peerings (SDPs) unless otherwise indicated these are vlans 1780-1783 Planned peerings (SDPs) Planned federated networks (GOLEs under single administrative mgmt)

  6. NSI Deployment NORDUnet Nordic infrastructure for Research & Education • NSI v1 deployed in 2011 • NSI v2 deploying now – Demos at SC 2012 Nov 2012 • NSI v2 Network Engineering Workshop – TIP 2013 Jan 2013 • To Date: – 250,000 NSI Connections set up in 2012 over the Automated GOLE fabric

  7. Experiment to Production NORDUnet Nordic infrastructure for Research & Education • The NSI architecture is proven • The [v1] Software is reliable – 250,000 Circuits (!) (Jeroen vdH) • v2 is happening… • The ancillary services such as performance monitoring, topo distribution need definition and community support to bring forward • Time for [early] Production deployments – If you are not participating in the trails, it is difficult to influence the evolution of the operational services – This is an opportunity for the entire R&E community to migrate in unison to a common globally deployed service – NORDUnet, NetherLight, StarLight are diving in…

  8. Evolution NORDUnet Nordic infrastructure for Research & Education • We want the AutoGOLE to continue – it’s a unique proving ground! • We want it to grow – to continue to be a rich facility for proving GLIF concepts and recommendations • I would suggest the GLIF community renew the commitment to the facility: – Add additional links to the fabric (mesh path finding, robustness, new multi-path technologies,…) – Increase the existing capacity in order to allow experiments to actually gain insights at scale – Add additional GOLEs and Networks to the fabric

  9. Leardership Transition NORDUnet Nordic infrastructure for Research & Education • Jerry Sobieski will be taking on important roles in GN3+ in 2013… • Need new Automated GOLE Chair • If you are interested, and like pain and frustration, but have a penchant for global domination, please let me know • The silver lining: The best network hardware and software engineers in the world are part of the team you get to hang with…

Recommend


More recommend