Hi Brian, to me, the main document describing segment routing is draft-ietf-spring-segment-routing. Within this document we describe the architecture and main functions.
draft-spring-segment-routing-mpls describes its instantiation to the mpls dataplane and draft-previdi-6man-segment-routung-header describes its ipv6 instantiation. Therefore, wouldn't be ok to have draft-previdi listed as informative in the segment routing architecture draft? Thanks. s. On Jul 24, 2015, at 10:49 AM, Brian E Carpenter <[email protected]> wrote: > Hi, > > It seems very strange to me that [I-D.previdi-6man-segment-routing-header] > is an Informative reference. Surely it is Normative (i.e., required reading > for an implementer)? > > That leads to another comment. As we know, if 6man accepts that I-D, it will > certainly be on the basis that the header is strictly confined to a > "consenting adults" domain (e.g. without middleboxes that can't deal with > strange extension headers, without MTU size & fragmentation problems, etc.). > But in draft-ietf-spring-segment-routing, there seems to be no definition > of a segment-routing domain, what happens at its boundaries, etc. The word > "domain" is used but without clear definition. I believe that is of normative > importance (i.e. it's not a use case issue) and is needed here so that > draft-previdi- can build on it. > > Brian > > -------------------------------------------------------------------- > IETF IPv6 working group mailing list > [email protected] > Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 > -------------------------------------------------------------------- _______________________________________________ spring mailing list [email protected] https://www.ietf.org/mailman/listinfo/spring
