Re: Comments on raft-fz-6man-ipv6-alt-mark-01

> On 1 Nov 2019, at 18:57, Tom Herbert <tom@herbertland.com> wrote:
> 
> On Fri, Nov 1, 2019 at 10:46 AM Ole Troan <otroan@employees.org> wrote:
>> 
>> DestOpt => measurement only  by node in DA
>> HBH => every router on the path with feature enabled
>> SRH TLV => every node along the SR path
>> DestOpt + SRH => every node along the SR path
>> 
> Ole,
> 
> So the last two have identical effects, hence we've introduced the
> complexity or redundant protocol mechanisms.

Yes, that should guide us in making choices here. 

Ole

> 
> Tom
> 
>> Cheers
>> Ole
>> 
>>>> On 1 Nov 2019, at 18:33, Tom Herbert <tom@herbertland.com> wrote:
>>> 
>>> Hello,
>>> 
>>>> From the draft: "Regarding Hop-By-Hop Options Header, if we consider
>>> its real deployment, it is sometimes dropped by legacy devices and not
>>> so used by intermediate nodes.  Destination Options Header is
>>> preferred."
>>> 
>>> I don't think this is helpful guidance. First of all, it's not just
>>> Hop-by-Hop options that can be dropped, it's pretty much packets with
>>> any extension heades or atypical protocols that might be dropped by
>>> legacy devices-- including packets with Destination Options or Routing
>>> Headers. Neither does it make sense that Destination Options Header is
>>> preferred, as correctly stated in the previous paragraph DO and HBH
>>> address difference use cases (i.e. DO is end to end, and HBH is per
>>> hop). Saying that DO is preferred is equivalent to saying that
>>> end-to-end performance measurements are preferred which I doubt is the
>>> intent. IMO, this whole paragraph could be removed without loss of
>>> content.
>>> 
>>> "SRH TLV can also be a good choice from this point of view.  The
>>> intermediated nodes that are not in the SID list can consider the SRH
>>> as a green field, they cannot support and bypass or support and dig
>>> into the SRH TLV."
>>> 
>>> I disagree with the conclusion that SRH TLV is a good choice. The
>>> implicit assumption in this paragraph is that somehow SRH EH is less
>>> likely to be dropped by intermediate nodes than other EH like DO and
>>> HBH. I don't think there's any data to support that. Additionally,
>>> it's not clear what use case an SRH TLV addresses that can't already
>>> be addressed by Destination or HBH options.
>>> 
>>> Tom
>>> 
>>> --------------------------------------------------------------------
>>> IETF IPv6 working group mailing list
>>> ipv6@ietf.org
>>> Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
>>> --------------------------------------------------------------------