Re: administrivia (on avoiding injury)

[Brian E Carpenter <brian@hursley.ibm.com>]

Joe, I don't agree that my text is a dilution. It's additional
and it is written very precisely to achive a very precise
effect. The two requirements are complementary, not mutually
exclusive.

My requirement is not for transparency or for sessions to proceed
after a glitch. It's for the ability to open *new* sessions from hosts 
that have *not* been modified from RFC 2460 conformance. That's different, 
and essential.

    Brian

Joe Abley wrote:
> 
> On Tue, Apr 10, 2001 at 02:24:54PM -0400, Margaret Wasserman wrote:
> > When are we expecting an updated requirements document that
> > reflects the extensive feedback sent to the list?
> 
> I have seen extensive feedback, but I'm not sure I've seen too much
> feedback that would require a change to the existing draft. I may
> have missed a thread or two.
> 
> > Also, how are we going to reach a resolution on our requirements
> > for IPv6 backwards compatibility, if any?  I have heard one or
> > two loud (and repeated) objections to Brian's proposed wording:
> >
> > "An IPv6 host running RFC 2460 IPv6, and running TCP and UDP
> > applications, including IPv4 applications running over
> > bump-in-the-stack or bump-in-the-API, must be able to open *new*
> > TCP and UDP sessions, regardless of which of the site's IPv6
> > provider links is up or down."
> >
> > I have also heard several people agree with this position.  I,
> > myself, strongly agree with this position.
> 
> The existing draft says this:
> 
>    Additionally, during failure events described above, multihoming
>    solutions must provide re-routing transparency for applications;
>    i.e. exchange of data between devices on the multi-homed network and
>    devices elsewhere on the Internet may proceed with no greater
>    interruption than transient packet loss during the re-routing event.
> 
> Brian's paragraph is a dilution of this requirement. While there has
> been discussion about specific approaches (e.g. the possibility of
> finding a solution at layers greater than three) I do not believe
> I have heard anybody disagree with the requirement as currently
> stated.
> 
> > Do we have rough consensus regarding this wording, or are there
> > other people who do not agree with this requirement?  I would
> > respectively submit that we could have rough consensus on this
> > point, despite Ohta-san's belief that this requirement is
> > unnecessary.
> 
> I would welcome comment on the paragraph in the existing draft.
> 
> Joe