Re: draft-wang-6man-flow-label-reflection

On 2014-11-14 18:52, Sheng Jiang wrote:
> Hi, Jeroen,
> 
> We can certainly add some consideration in the future version regarding
> to ICMPv6 and "Packet too big".

Without it, it is quite futile to document more about Flow Labels...

> However, could you be more specific what your concern is regarding to ICMPv6 and flow label reflection?

See the below mentioned draft and the whole thread about Flow Labels...

> BTW, the v6ops draft you mentioned does not have any discussion regarding to flow label.

The word "Flow Label" is hidden in the ECMP calculation.

Greets,
 Jeroen

> Regards,
> 
> Sheng
> ________________________________________
> From: ipv6 [ipv6-bounces@ietf.org] on behalf of Jeroen Massar [jeroen@massar.ch]
> Sent: 14 November 2014 16:43
> To: 6man WG
> Subject: draft-wang-6man-flow-label-reflection
> 
> As the words ICMPv6 nor "Packet Too Big" are in this draft it already
> has big problems.
> 
> ICMPv6 is part of a "Flow" as it signals errors related that flow.
> 
> Hence, not having a mention of them makes that draft not work.
> 
> 
> Further see:
>  http://datatracker.ietf.org/doc/draft-v6ops-pmtud-ecmp-problem/
> 
> For a writeup why the Flow Label is broken.
> 
> Note also that as "Flow Label" is supposed to be used for Load
> Balancing, but I have not found a hardware device yet that is actually
> using it for that purpose, they just look at src + dst and optionally
> ports if they want to go that low...
> 
> As stated elsewhere: the 20 bits of the Flow Label are a waste of header
> space.
> 
> Greets,
>  Jeroen
> 
> --------------------------------------------------------------------
> IETF IPv6 working group mailing list
> ipv6@ietf.org
> Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
> --------------------------------------------------------------------
> 