Speaking as WG member: Hi Tony, Thank you for starting this discussion. I’ve reviewed the draft and have a few comments and questions.
1. I think the concept of a flood reflection cluster should be defined earlier in the document as opposed to being introduced in the TLV description. It seems there can only be one cluster per level-1 area so why isn’t the cluster implied by the area? I don’t see the reason for having a separate Flood Reduction Cluster ID? 1. On Page 7, there is the statement “A solution without tunnels…” I think this should be removed or at least there should be a qualification that it is beyond the scope of the document. I know there are subtle allusions to the tunnel-less solution in other sections but these introduce confusion as well. 1. For the TLVs in section 3, 4, and 5, it is a “MUST” to only have a single instance of the TLVs. However, if there are more than one, this violation is ignored and the first instance of the TLV is used. Perhaps, this should be a “SHOULD”. Also, these violations SHOULD be logged subject to rate-limiting. 1. Section 6 – How does a flood reflector ensure that there are no normal L2 adjacencies? Will adjacency establishment fail if an attempt is made to establish one? This should be specified. 1. Section 6 – Establishment of tunnels between flood reflector clients is a MAY yet the solution work without it unless all the L2 routes are leaked into L1. This is related to #2. 1. Section 7 – This section needs to specify what happens if the adjacency criteria are not met. 2. Section 8 – Does this section assume no tunnels? Is the assumption that if there is an L1/L2 edge router that is not a flood reflector client, there will be another L1/L2 edge that is? Or is this to cover cases where there is an existing portion of the L2 topology that doesn’t support flood reflection. If it is the latter case, than leaking is required to get the L2 routes for this portion of the L2 topology to the flood reflector(s). 1. I don’t understand the last statement in section 9. What do you mean by “look ‘shorter’”? Why wouldn’t the flood reflected path cost be accurate? The L2-only path would certainly be accurate. Thanks, Acee From: Lsr <[email protected]> on behalf of Tony Przygienda <[email protected]> Date: Friday, July 9, 2021 at 1:27 AM To: lsr <[email protected]> Subject: [Lsr] Request to consider Flood Reflection going into LC Dear chairs, flood reflection is implemented/shipped and deploying and we’re on early alloc codepoints that expire start Aug’. No further discussions on the list happened. As far as we see as authors stuff looks shaken out, all things that were found in implementation/use are in the latest draft version and I’d like to test the waters of LC’ing it rather than dragging it on as draft and asking for temporary alloc again which does not seem to serve a purpose AFAIS. thanks -- tony
_______________________________________________ Lsr mailing list [email protected] https://www.ietf.org/mailman/listinfo/lsr
