Re: [bess] WG Last Call for draft-ietf-bess-evpn-prefix-advertisement-04

[Eric C Rosen <[e-mail]>]

While I would like to see this document advance eventually, I don't 
think it is ready yet.

Main points:

- There is no clear explanation of the key concept of "overlay index".  
In particular, there is no real explanation of when to use an overlay 
index, or of when to use each kind of overlay index. There are some use 
case descriptions, and some examples of the sort "in this use case use 
this kind of overlay index", but no rules that specify the precise 
circumstances under which it is appropriate to use each kind of overlay 
index.

- There are no rules given that say how an NVE knows whether to 
originate an RT-5 route, or knows how to construct an RT-5 route.  A 
number of use cases are walked through, which is helpful, but that is 
not a substitute for a real specification.

- In the discussion of use case, there are statements like "support of 
this use case is REQUIRED".  But it is very difficult to know exactly 
which features of the protocol are being mandated.

- Too much of the draft is the "RT-5's are good" sale pitch, which is 
repeated three times.  (Sections 2.2, 4, and 6.)  A single time would do.

- The talk of IP-VRFs is a bit misleading, as this might be taken to 
suggest that the document provides a way to interoperate L3VPN with EVPN.

I've attached the draft with some more comments in-line, look for lines 
beginning with ****.