Folks,
I think that we have agreed to the following:
- We should include a reference to S5.3 of RFC 3168, reminding implementers
that when reassembling a packet, if any packet has ECN set, the reassembled
packet MUST have ECN set. While this may be orthogonal to the draft, it is
included as a courtesy.
- We don't need to add any text about copying DSCP or ECN bits back and forth
between the payload and delivery headers, as this issues are beyond the scope
of the document.
Have we in fact agreed on this?
Ron
> -----Original Message-----
> From: Eggert, Lars [mailto:l...@netapp.com]
> Sent: Thursday, August 08, 2013 3:26 AM
> To: Carlos Pignataro (cpignata)
> Cc: Ronald Bonica; Bob Briscoe; Internet Area
> Subject: Re: [Int-area] draft-bonica-intarea-gre-mtu and ECN
>
> Hi,
>
> On Aug 7, 2013, at 22:18, Carlos Pignataro (cpignata)
> <cpign...@cisco.com> wrote:
> > In other words, noting S5.3 of RFC 3168 in draft-bonica-intarea-gre-
> mtu does not add as compared to not noting it -- the requirement
> already exists, and is applicable to any frag/reassembly, including
> GRE, any other tunnel, or any other protocol, as the ECN considerations
> at the IP layer. RFC 3168 specifies transport considerations for TCP
> and not for GRE also.
>
> you're correct. In practice, however, not everyone is familiar with the
> contents of all related RFCs when implementing a given protocol
> extension, and I think pointing implementors at relevant related specs
> - esp. when mandatory to support - is a courtesy.
>
> > Net-net, in my humble opinion, since draft-bonica-intarea-gre-mtu
> concerns itself with fragmentation strategies for GRE, these two ECN
> issues (ECN at reassembly and tunnels) seem orthogonal to the scope and
> goal of the document, and better dealt with in a separate place instead
> of intertwining them in draft-bonica-intarea-gre-mtu.
>
> Which document specifies the reassembly of fragments carried in GRE?
>
> Lars
_______________________________________________
Int-area mailing list
Int-area@ietf.org
https://www.ietf.org/mailman/listinfo/int-area
