Re: draft-pillay-esnault-ospf-flooding-05.txt

[Pekka Savola <pekkas@netcore.fi>]

On Thu, 12 Jun 2003, Bill Fenner wrote:
> >If flooding is really considered to be a problem, my argument would be 
> >that one needs to at least consider why re-using RFC1793 for this specific 
> >purpose is the *right* approach.  I'm not really sure of that, but sounds 
> >like a spec abuse (and who knows how many assumptions in RFC1793 may 
> >become invalid when it's applied to a scenario it wasn't meant for?)
> 
>   This document was reviewed by (submitted by!) the OSPF WG, and in
> particular was reviewed by the author of RFC 1793 while he was still
> participating in the working group.  Also, Alex has specific OSPF
> implementation experience and has reviewed this spec.

Ok, it seems to have had enough protocol expertise.
 
>   I realize that we don't like re-use abuse, but I think that this is
> not just a "cute hack" but also a reasonable, low-barrier solution to
> something that's not necessarily everybody's problem.  If you don't
> agree that it's a problem, then why do you want a heavier weight
> solution?

One might make an observation that "excessive" flooding started to become
a problem once more and more junk like TE capabilities were put into OSPF.  
Do we discourage against putting more junk in the protocol?  Did we forget
to tell it plainly enough that adding more junk in the protocol will have
consequences (like more control data transferred), and one has to live
with that?  Instead, we complicate the protocol further and make it less
robust to deal with the issue of putting more junk in the protocol (by
putting more junk in the protocol).  And this is not a problem of OSPF
alone, unfortunately...

But I digress...

It would seem to me that a subset of RFC1793 (rather than all of it) would
also be enough. An alternative operational approach might be to just
manually be able to set the refresh timeout higher.

But as this is just an informational document, I see little use spending 
too many cycles on it.  Do whatever you feel appropriate.

btw. I also note that the just-revised version includes the IPR statement, 
making this doubly useless.  Oh well.

-- 
Pekka Savola                 "You each name yourselves king, yet the
Netcore Oy                    kingdom bleeds."
Systems. Networks. Security. -- George R.R. Martin: A Clash of Kings