Re: Reopening jumbo frames in IS-IS

[Eric Rosen <erosen at cisco.com>]

> they were not going to specify how jumboframes work at all
> (nevermind with IS-IS) and please do not that that answer to mean that
> the IETF should do so

Really, with regard to jumbo frames  the marketplace has spoken and the IEEE
has lost. Jumbo frames are a  defacto standard.  (Of course, just as happens
in theIETF,  the IEEE folks don't always  admit it when they've  lost out to
the marketplace.)

Given that  jumbo frames are  in use,  it seems silly  to say that  we can't
specify how to use them for IP protocols.  There is even some precedent; RFC
3032 (MPLS encaps)  mentions the possibility of using  ethernet frames which
contain 1504 or 1508 bytes.  

It's even sillier to say that (a) we're not allowed to do a carefully scoped
standard that solves a well-specified and immediate problem in the use of an
existing de  facto standard, but that (b)  it's fine to do  a standard which
tells folks to ignore IEEE entirely and do bridging in a completely new way.