Re: Objection to draft-ietf-6man-rfc4291bis-07.txt

[james woodyatt <[e-mail]>]

On Feb 23, 2017, at 05:40, Peter Hessler <[e-mail]> wrote:
> 
> Restricting all subnets to The One True Size(tm) of /64 is utterly
> ridiculous.  Sure, that may be an artificial limitation of SLAAC and
> various other technologies, but *those* can have limitations.
> 
> Limiting it inside the entire specification is even stupider of an idea
> than still supporting Classful networks.
> […]

It would help if those objecting to the promotion of RFC 4291 to Standard, unless the requirement for subnet prefixes to be generally /64 (except where noted by standards track documents), would please remember that SLAAC is only one of several technologies dependent on it. That’s why this draft now includes a reference to RFC 7421, which lists a non-exhaustive list of several things that are broken on subnets where prefixes longer than /64 are used.

My counter to these objections should be regarded from an overview perspective.

IPv6 is about more than just routers and the few thousands of network operators who love them. It’s also about hosts, their application software and the billions of ordinary people who use them.

Many of those Proposed Standard protocols listed in RFC 7421 (and a few others that were overlooked in that document, but which are nevertheless important) may not seem strictly necessary to the operator community bringing these objections, but they are necessary to make host applications provide user experiences on which billions of people depend. Insisting that RFC 4291 drop this requirement to be published as a full Standard is tantamount to insisting that IPv6 itself not be required to meet the needs of billions of humans around the world because Classful Networks make a few thousand network operators have a sad.

Needs of the many. Needs of the few. Seems like an obvious balancing problem to me.


--james woodyatt <[e-mail]>