Towards a Disjoint IRR

In message <199505011712.AA18161 at cat.isi.edu>, Cengiz Alaettinoglu writes:
> 
> Daniel Karrenberg (Daniel.Karrenberg at ripe.net) on May 1:
> > Q2) Should users be encouraged to register objects in multiple RRs.
> > 
> > The answer is NO because
> > 
> > 	- it creates inconsistency problems
> 
> Daniel,
> 
> I think Curtis's point was that the inconsistency is not actually a
> problem because everyone will going to thrust its own data first. Curtis
> please correct me if I am wrong.

See my last two messages.  We with "thrust" our own aut-num, inet-rtr,
direct attached customers, and do secondary for other registries that
are authoritative for other parts of the topology so our config
process is not affected by occasional failure to reach another part of
the Internet.

> I think inconsistency is a problem, especially if the conflict is
> happening in two sources which are not in my control (say I am the RA
> and the conflict is between MCI.db and CANet.db, who should I trust).

If everyone is keeping all the data and considering their own copy
authoritative, I agree that there would be a big consistency problem.

Sorry if I was unclear.

> Cengiz
> 
> -- 
> Cengiz Alaettinoglu       Information Sciences Institute
> (310) 822-1511            University of Southern California
> http://www.isi.edu/div7/people/cengiz.home

Curtis



-------- Logged at  Mon May 1 22:59:21 MET DST 1995  ---------