On 18 Nov 2010, at 16:13 , Suresh Krishnan wrote: > You said that, in your implementation,
I don't have an implementation. I'm an independent consultant, primarily to folks who happen to operate some large IP networks. In that position, I have NDAs in place not only with my clients, but also with some organisations that work with my clients. > it is easier for the hardware to parse a hop-by-hop or > destination options rather than an extension header. That is NOT what I said. What I said was that: MULTIPLE vendors have already shipped and deployed routers that already contain silicon that can parse/ parse past IPv6 optional headers in hardware at wire-speed. The key word above is: MULTIPLE My understanding is that routers with these capabilities in the silicon are already deployed in multiple large IP networks, and that some operational networks rely on that capability (e.g. to be able to apply ACL rules based on transport-layer information, such as transport protocol, transport port numbers). Adding any extension header, other than an option carried within the existing IPv6 Destination Options header or the IPv6 Hop-by-Hop Options header, breaks that important operational capability. We all ought to try really hard to avoid breaking deployed standards-compliant IPv6 networks. Yours, Ran -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------