> On Thu, Jul 17, 2014 at 6:27 AM, Ken Gray (kegray) <[email protected]> wrote:
>> Where Dino and I agree, and not to pick on Uri - I've seen the statement
>> "additional headers/metadata" a few times now and it STILL sounds like
>> "work around SFC", particularly in respect to the metadata part (since
>> passing metadata was one of SFC's points).
>> 
>> If you want to tilt at that windmill, I don't think that binding the
>> solution to a particular transport encap is the optimal way to do it (if
>> you can avoid doing so).
>> 
> By this line of thinking TCP options should be moved to extension
> headers. If the meta data is part of the transport protocol and needed
> for its operation then it's should be within the transport header not
> buried at some other place in the packet. There's already a lot of
> precedent for this in IETF protocols and we had more than thirty years
> of operational experience dealing with optional data in several core
> protocols. I really don't see that the encapsulation protocol case is
> so unique that a new model needs to be adopted.

Well stated and 100% agree.

Dino

_______________________________________________
nvo3 mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/nvo3

Reply via email to