60 likes | 184 Vues
This document discusses two drafts focusing on router convergence benchmarking: draft-papneja-bgp-basic-dp-convergence-02 and draft-varlashkin-router-conv-bench-00. The former addresses data-plane only BGP convergence, while the latter outlines methodologies that complement it. The goal is to align reality with users' expectations and refine the benchmarking process for routers. Input is welcome as the working group aims for readiness by the next IETF meeting, addressing topology and failure scenarios to ensure robust performance evaluation.
E N D
Routers convergence benchmarking draft-papneja-bgp-basic-dp-convergence-02 draft-varlashkin-router-conv-bench-00 Ilya Varlashkin ilya@easynet.com Rajiv Papneja rajiv.papneja@huawei.com Bhavani Parise bhavani@cisco.com Dean Leedlee@ixiacom.com Sue Hares skh@ndzh.com
Current status • draft-papneja-bgp-basic-dp-convergence-02 • data-plane only • assumes fwd ability implies BGP convergence • draft-varlashkin-router-conv-bench-00 • mentioned @IETF81, now published • works with independent data- and control-planes • The two drafts complement each other Routers convergence benchmarking
Two drafts side-by-side Routers convergence benchmarking
Work roadmap • Reality alignment • ensure methods deliver values that people want to know • assert feasibility • Fine-tune methodology • eliminate (most of) dependency on non-DUT • define equipment config in generic terms • Sort out terminology Routers convergence benchmarking
Test topology R1 R3 1 1 9 C2 [S] M1 DUT NetA C1 2 3 R2 R4 ER1 ERn NetBn NetB1 Routers convergence benchmarking
TODO list • Agree on approach • Do we need to change test topology? • Do we need to change failure scenarios? • Input welcome! • Aim for WG-item readiness by next IETF • Consider need for amending RFC4098 Routers convergence benchmarking