Re: FW: [Fwd: Re: [MOBILE-IP] GRE mailing list]

[Rene Tio <tor@redback.com>]

Comments inline.

> -----Original Message-----
> From: charliep@iprg.nokia.com [mailto:charliep@iprg.nokia.com]
> Sent: Friday, March 10, 2000 2:54 PM
> To: Randy Bush; David Oran
> Subject: Re: [Fwd: Re: [MOBILE-IP] GRE mailing list]
> 
> 
> 
> Hello Dave and Randy,
> 
> Would one of you folks be willing to forward my message to
> the mailing list?  If I subscribe to another mailing list,
> my pistons will melt.
> 
> Regards,
> Charlie P.
> 
> 
> Randy Bush wrote:
> >
> > only subscribers may post
> >
> > > Date: Fri, 10 Mar 2000 08:25:15 -0800
> > > From: "Charles E. Perkins" <charliep@iprg.nokia.com>
> > > Organization: Nokia Research Center
> > > To: gre@ops.ietf.org
> > > CC: David Oran <oran@cisco.com>
> > > Subject: [Fwd: Re: [MOBILE-IP] GRE mailing list]
> > >
> > >
> > > Hello folks,
> > >
> > > My experience with GRE is that it has a lot of features that
> > > would not be needed for Mobile IP.  While I have not followed
> > > the recent discussion, I note that existing GRE deployments
> > > would probably _not_ handle any extensions.

Given my ignorance of Mobile IP, I would have to as what extensions you need
and whether they covered by Gopal's draft.  RFC 1701 GRE has a lot of
features which were of dubious use anyway, but all of them are (a) easy to
understand (b) optional (compare this to L2TP).  Basic GRE itself (8 bytes)
however has been in use for a long time.  I would worry about RFC 2004
contributing to protocol bloat if it doesn't give you any benefit other than
saving 20 bytes of outer IP header.

> > > Thus, if needed functionality is going to be acquired by
> > > new extensions, new GRE deployments would be needed.  So,
> > > why shouldn't Mobile IP define the necessary tunneling
> > > protocol that we need?

Because deploying a new protocol would force you into a software upgrade,
just as deploying a new version of GRE would.  Besides, GRE has widespread
acceptance and deployment as compared to RFC 2004 (as you noted below).

> > > I also note that the existing Minimal Encapsulation protocol
> > > (RFC 2004) is a likely candidate revision to suit any additional
> > > tunneling needs for Mobile IP.  There are reserved bits we
> > > could use to control new features, and the existing deployment
> > > of RFC 2004 is so small that obviously people don't like it
> > > much anyway -- so it needs to be changed to be useful.
> > >
> > > I would rather implement a small tunneling protocol that does
> > > exactly what Mobile IP needs, instead of a new protocol from
> > > a big specification.  Is GREng likely to be a big one, or a
> > > small one?

GRE has never been big, depending on your definition.  Does 20 bytes matter
that much in mobile IP applications or does "big" refer to complexity?

Rene

> > > Regards,
> > > Charlie P.
> > >
> > >
> > > Dave Oran wrote:
> > > >
> > > > In case any of you have missed it, there is now a dedicated mailing
> list to
> > > > discuss possbile future enhancements to and progression beyond
> proposed
> > > > standard for GRE.
> > > >
> > > > The list is
> > > >         gre@ops.ietf.org
> > > >
> > > > subscribe via mail to gre-request@ops.ietf.org
> > > >
> > > > Please join in there so that we can get all the various Working Group
> > > > requirements and interests exposed in one place.
> > > >
> > > > Thanks, Dave Oran
> > >
> 
> 
>