This document makes several erroneous assumptions, IMO:

- 2.2: applications don't "tailor the MSS" (transport layers do)

- overall: PMTUD is deprecated, in favor of PLMTUD

-overall: in a tunneling system, the packetization layer is the tunnel
encapsulator, which means that layer should be doing MTU discovery

- 3.1: PTB ICMPs are generated when the packet size is greater than the
link MTU, not the other way around

- 3.1: IPv6 doesn't have a DF bit

- the overall approach misuses DF and PTB messages, IMO

These issues have been discussed repeatedly on the INTAREA mailing list.

Joe

On 6/30/2015 7:57 AM, Saumya Dikshit (sadikshi) wrote:
> Hello All,
> 
> We have submitted a draft related to Path MTU Discovery in an end to end
> path,
> between devices connected over a Vxlan network. The draft outlines existing 
> Problems and a potential solution to go around them, without tinkering with
> existing standards.
> 
> https://datatracker.ietf.org/doc/draft-saum-nvo3-pmtud-over-vxlan/
> 
> Please provide your comments and suggestions on the same.
> 
> Thanks and Regards,
> Saumya.
> 
> 
> 
> _______________________________________________
> nvo3 mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/nvo3
> 

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

Reply via email to