From: Howard C. Berkowitz (hcb@gettcomm.com)
Date: Mon Oct 13 2003 - 16:31:27 GMT-3
At 7:08 PM +0100 10/13/03, Stuart Reabow wrote:
>Hi all,
>
>I've been working through numerous labs that have multiple routing protocols
>per router, sometimes with 1 or 2 proptocols on all the routers. If this is
>the case, are we supposed to have all routes in all protocols, or can we
>choose one or two protocols to contain all the routes & effectivity ignore
>the others? If this gives full connectivity without loops is this wrong?
>Also which would be the better one's to use, OSPF, BGP, EIGRP etc....
I'm afraid I'm answering on a real-world, not lab basis, but here it
goes. In general, it is best to minimize the information given to any
IGP. Within an AS, iBGP should generally have the most information
assuming there are large numbers of routes.
With respect to BGP and the global Internet, it's considered
responsible best current practice for ISPs to minimize the number of
routes they announce. There's a current thread on NANOG that actually
has an unstated assumption: there can be massive route injection
reductions if ISPs bilaterally agree on aggregate address blocks to
be given to their joint multihomed customers. Realistically, that
makes sense only if the joint ISPs send each other full routes,
potentially including customer routes. There are technical tradeoffs
and proprietary data issues in how much of an internal view is shared.
***Get your CCIE and a FREE vacation: Shop.GroupStudy.com***
This archive was generated by hypermail 2.1.4 : Mon Nov 24 2003 - 07:53:01 GMT-3