1 / 11

Using E2E technology for LHC

Using E2E technology for LHC. Apr 3, 2006 HEPiX Spring Meeting 2006 marco.marletta@garr.it. Agenda. Multilayer and multidomain networks End-to-end or edge-to-edge? A case study: KarBol Challenges The LHCOPN. Multilayer Research Networks. Research networks multilayer approach LAN MAN

brian
Télécharger la présentation

Using E2E technology for LHC

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. Using E2E technology for LHC Apr 3, 2006 HEPiX Spring Meeting 2006 marco.marletta@garr.it

  2. Agenda • Multilayer and multidomain networks • End-to-end or edge-to-edge? • A case study: KarBol • Challenges • The LHCOPN [HEPiX Spring meeting 2006]

  3. Multilayer Research Networks • Research networks multilayer approach • LAN • MAN • Regional Network • NREN • Continental backbones (GÉANT) • Multiple domains • Multiple responsibilities [HEPiX Spring meeting 2006]

  4. End-to-end or edge-to-edge? • IP was designed as an end-to-end L3 protocol • In principle, “lightpaths” use suggests a clear L1 or L2 end-to-end channel • L2 services also possible using MPLS-L2VPNs and VPLS • However, often it is necessary to cross several edge borders • “stitching” of services across domain borders is necessary • Sharing the same setup (MTU, buffer and queue sizes) • Otherwise performance is affected! • “stitching” of people is crucial [HEPiX Spring meeting 2006]

  5. G-WiN GARR DFN-PE FZK-CE GARR-PE CNAF-CE Geant A case study: KarBol • In 2004, an MPLS-L2VPN was created between INFN-CNAF Bologna (Italy) and FZK-Karlsruhe (Germany) • A model to implement end-to-end services between tier centers • L2 connectivity between the 2 T1 centers obtained across an L3 path • 5 different competence domains involved • New technical challenges faced and solved [HEPiX Spring meeting 2006]

  6. Challenges • Multidomain L2VPN: something router manufacturers did not consider enough • Interoperability between platforms • An L2 path can be a security backdoor into someone’s LAN • An alternative using L3VPN was studied • A trusted relationship is needed • Sharing L3 addresses (private, public?) • Complex setup • VPLS could be the solution to scale to T2 numbers [HEPiX Spring meeting 2006]

  7. More challenges: CBF • Cross-Border-Fibre, on rent or IRU, or digged (!) • DWDM equipments, 10G lambdas • Complementing GÉANT2 infrastructure • Offering reciprocal backup path from T1 to T0 SWITCHCH DFNDE GARRIT CNAFBologna FZKKarlsruhe [HEPiX Spring meeting 2006]

  8. LHCOPN – LHC Optical Private Network • Every Tier1 will be connected to the Tier0 with a direct 10Gbps “lightpath”. • Those lightpaths will be of different kinds: • single or concatenated layer 1 links (STM64, LANPHY, WANPHY) • layer 2 VLANs • Tier1s should also provide a dedicated backup link to Tier0 • during the startup phase (Service challenges) backup will be provided via routed paths (GN2, Esnet, Abilene, NRENs..). [HEPiX Spring meeting 2006]

  9. LHCOPN challenges • Security • The LHCOPN will bypass the security system (firewall, IDS...) already in place at every Tier; current technologies cannot deal with the requested bandwidth • Operations The ENOC is Network Co-ordination Service. It is required to: • look after network issues for EGEE and LCG • receive network TTS from NRENs, analyse them and provide relevant information to the GGUS who will then interact with the users • monitor the e2e status of the lightpaths and trigger the appropriate corrective actions • Monitoring • Several metrics, several different devices, and several OSI stack levels to monitor [HEPiX Spring meeting 2006]

  10. Current LHCOPN topology [HEPiX Spring meeting 2006]

  11. The End Thanks to Edoardo Martelli, CERN

More Related