1 / 13

Workshop Introduction - Finding the Match

Workshop Introduction - Finding the Match. Lars Fischer LHCONE Workshop CERN, 13-14 December 2012. Welcome. Welcome to CERN Thanks to CERN for hosting Thanks to all of you for attending A community workshop Part of the LHCONE process Expanded LHCONE working group

yestin
Download Presentation

Workshop Introduction - Finding the Match

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Workshop Introduction-Finding the Match Lars FischerLHCONE Workshop CERN, 13-14 December 2012

  2. Welcome • Welcome to CERN • Thanks to CERN for hosting • Thanks to all of you for attending • A community workshop • Part of the LHCONE process • Expanded LHCONE working group • A meeting of networkers and WLCG participants • This is a workshop • We’re here to jointly find a way forward

  3. Why LHC networking? • The dual challenge • Predictable performance • Protecting everyone else from overload • Why LHCOPN was successful • Dedicated, hierarchical data distribution network for limited number of sites • Why do we need LHCONE • WLCG computation model has changed • Any T2/T3 to/from any T1 traffic • … and why can’t we reuse the LHCONE model • Larger number of sites • Changing traffic patterns

  4. LHCONE Activities “Interface with the LHC software stacks” 6. ESnet/USLHCNet 2. SARA/ SURFnet 1. DANTE 3. Internet2 4. NORDUnet SURFnet 5. Internet2/ DANTE “Carrier Ethernet/ OAM/ TRILL” “VRF” “Diagnostics” “P2P Service” “OpenFlow” Innovation Fix it Result of the LHCONE meeting in Amsterdam, December 1&2, 2011

  5. What has been achieved? • VRF solution deployed • “LHCONE 1.0” moved to production • Sites getting connected, traffic growing • Progress on diagnostics, monitoring • Engineering focus moved to “LHCONE 2.0” • Progress on service definitions • Discussion, better understanding of next-generation technology options

  6. Point-to-Point Status • Consensus among networking people • What can be offered as a service • Understanding of how a service can be managed across domains • Service Definition in place, documented • Document extensivelydiscussed, rough consensusestablished • Point-to-Point protocols • NSI frameworkagreed as the longterm solution • Both NSI and IDCP in deployment, canbeused for LHC trafficnow • Major networksmovingtowards NSI2.0 • Pilot discussed, not yetinitiated • Service not yet in operation

  7. Plenty of progress elsewhere • ESnet and Internet2 BoD deployments • Extensive deployment across North America, strong experience with managing a BoD service • European BoD deployment • GÉANT, growing number of NRENs • Commitment to have NSI2.0 deployed by Q1 2014 • NSI 2.0 framework • Connection Service standardization • AutoGOLEtestbed • Work in Progress • How do we leverage these to bring a service to the LHCONE?

  8. Point-to-Point challenge • Point-to-point require experiments to approach networking differently • Not a drop-in replacement for the existing network service. • Network transport is unchanged • Network resource usage is no longer transparent to the application workflow • Point-to-Point reservation must be integrated with applications • Must be part of workflow design • For WCLG, this means • Changing a complex set of applications • Finding the right match between PtP technology and the workflow of the experiments

  9. What makes P2P attractive? • Easier to manage peers • Cleaner way to manage resource allocation • Enabling use of available resources • Ensuring resources are used as intended • Cleaner way to handle the firewall issues • Avoid VRF design kludge and administration issues

  10. PtP in WLCG workflow • We need to integrate point-to-point transport into the WLCG applications • We must understand where in the workflows used point-to-point circuits can be requested & terminated • We need to understand the full production cycle, with point-to-point part of the picture • Networkers need to understand the WLCG workflow • WLCG site and application people must understand what point-to-point can offer • Where we are: • We have the service definition… • … but we’re missing a clear picture of the system and the system elements where the service is deployed • …and agreement on the components we will use to build that system • We need a joint group to do that!

  11. Expectations for PtP • We hope to • Identify and agree simple use cases for the service, implement quickly • Gradually expand the service, build as one component of the comprehensive LHCONE network service • Point-to-point and VRF in parallel for some time (at least) • We (the networkers) need help • Including application experts, workflow managers, sites, users • Collaborative effort • We need The Big Picture™ • Working on the components will not ensure progress

  12. Workshop structure • Thursday: building understanding • Explain the basics, challenges, and status of point-to-point networking • Explain the basics, challenges, status, and changes of the WLCG applications and computation model • Friday: initiating design • Discuss how point-to-point can fit into WLCG model • Discuss the challenges posed to point-to-point by the WLCG applications • Form consensus on a realistic way forward • Form a group of people to carry the work forward - we need more than networkers for this to succeed

  13. What I hope to have tomorrow • A clear picture of how the point-to-point network model and the WLCG computation model match • A roadmap for what must be done to (eventually) deploy and use a service based on this match • A clear set of (simple) next steps for implementation in first half 2013 • An expanded LHCONE group – clear point of contact between networkers and experiments.

More Related