100 likes | 113 Views
LDP Data Plane Convergence Benchmarking. Thomas Eriksson, TeliaSonera Scott Poretsky, Quarry Rajiv Papneja, Isocore. Introduction. Service providers need fast convergence in their networks Standard for LDP data plane convergence benchmarking needed
E N D
LDP Data Plane Convergence Benchmarking Thomas Eriksson, TeliaSonera Scott Poretsky, Quarry Rajiv Papneja, Isocore
Introduction • Service providers need fast convergence in their networks • Standard for LDP data plane convergence benchmarking needed • We have presented the terminology earlier and today present some methodology for discussion IETF 62 Minneapolis
Terminology • No updates for this meeting • Currently at rev -02 • New terms will be added if needed for methodology IETF 62 Minneapolis
Methodology DUTs: ingress LSR, Core LSR and Egress LSR IETF 62 Minneapolis
Methodology Topology for single next hop FECs IETF 62 Minneapolis
Methodology Topology for multiple next hop FECs IETF 62 Minneapolis
Methodology • Multiple next hop FECs might be found in two cases • Two routers with parallel links • One router having ECMP routing from itself to a number of other routers • Test to be performed for both scenarios IETF 62 Minneapolis
Methodology • Tests to be performed for both good and bad news (E.g. up/down FEC convergence events) • It shall be possible to test for any routing protocol that provides FECs for LDP to bind labels. e.g. IGPs or IGPs with loop free alternates algorithm IETF 62 Minneapolis
Methodology • Interesting FEC convergence events • IGP metric change • IGP LSP/LSA with remote topology change • Lost IGP adjacency • Lost LDP session • Label withdrawal • Interface shutdown (local/remote) • Fiber pull (local/remote) • Removed IP address IETF 62 Minneapolis
Methodology • Next steps • Detailed test cases • Considerations for the tests • Write all this up and submit Comments? IETF 62 Minneapolis