On Thu, Jul 25, 2013 at 12:52 AM, Thomas Narten <[email protected]> wrote:

> Hi Lizhong.
>
> > After reading this draft, I find it still combines many possible options
> > (e.g, dataplane). Hope to see a draft with much more clear opinion in
> next
> > version.
>
> Not sure what you mean here when you say "combines many possible
> options (e.g., dataplane)."
>
[Lizhong] In section 10, I see at least three dataplane ecapsulation
approach (vxlan, nvgre, mpls over gre). I hope the WG will not standardize
so many kinds of encapsulations, otherwise it would be a nightmare for
vendors. The arch document should weight among the candidates, and gives
out the standardization direction, or at least should point out the reason
why we need three encapsulations.

>From the market point, I haven't heard of any potential NIC supporting MPLS
over GRE. Anyone heard of that?

Regards
Lizhong


> One thing that became apparently in writing this document is that the
> WG hasn't actually (on the record) decided number of things. To me,
> some of the decisions are obvious (in the sense that they make sense
> to me and others I've talked to), but that is not the same as the WG
> actually weighing in making a choice.
>
> One of the key differences between the architecture document and the
> framework document is that the framework is a bit more open ended in
> terms of possible approaches, whereas the architecture will reflect
> the architectural choices the WG has made for NVO3.
>
> So, an important goal in putting together this strawman is to get
> feedback and see where folk seem to say "yes", and where they say
> "actually, we should do something different..."
>
> I'll respond to your specific questions in separate threads, so that
> it's a bit easier to track/follow the issues.
>
> Thomas
>
>
_______________________________________________
nvo3 mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/nvo3

Reply via email to