From: EA Louie (elouie@xxxxxxxxx)
Date: Fri Nov 02 2001 - 00:48:16 GMT-3
actually, Dennis, the point I was making is that there's really no such
thing as IGP<-->eBGP redistribution (almost by definition). If you have an
IGP that is defined within a BGP AS, that makes the redistribution
IGP<-->iBGP by definition. The distinction between iBGP and eBGP is only
known within BGP. If you have an IGP that spans two different BGP AS's, it
still isn't eBGP<-->IGP redistribution. It's an iBGP<-->IGP redistribution.
If I'm off-base, I hope one of you will correct me. (Unless it's Dennis
;-)
-e-
----- Original Message -----
From: "Dennis Laganiere" <dennisl@advancedbionics.com>
To: "'EA Louie'" <elouie@yahoo.com>; <ccielab@groupstudy.com>
Sent: Thursday, November 01, 2001 8:46 PM
Subject: RE: redistribution consolidation - Now what?
> Your absolutely right about the BGP-to-BGP's. Since you can't have two
> instances running, you can't redistribute.
>
> I disagree about excluding IGP-to-BGP and BGP-to-IGP redistribution.
While
> it's not recommended in the real world, it can still be done, so we should
> know how to do it. By the way, one of the main reasons not to do this is
to
> avoid BGP learning a route from IGP that it learned from BGP... :-)
>
> --- Dennis
>
> -----Original Message-----
> From: EA Louie [mailto:elouie@yahoo.com]
> Sent: Thursday, November 01, 2001 4:58 PM
> To: Dennis Laganiere; ccielab@groupstudy.com
> Subject: Re: redistribution consolidation - Now what?
>
>
> > Wow... I've sent out over 50 copies of the redistribution matrix in
> > the last few hours. My keyboard is starting to heat up. :-)
> >
>
> I sent out quite a few, too... when it's 'done', I'll post it.
>
> comments within for the redistribution matrix
>
> I thought we covered these below already, which were all the IGP to IGP
> redistribution rules...
> > * OSPF to EIGRP
> > * IGRP to EIGRP
> > * EIGRP to IGRP
> > * EIGRP to EIGRP
> > * EIGRP to OSPF
> > * RIPv2 to EIGRP
>
> These don't really apply from a redistribution perspective - correct me if
> i'm wrong
> > * EIGRP to eBGP
> > * RIPv1 to eBGP
> > * RIPv2 to eBGP
> > * IGRP to eBGP
> > * eBGP to RIPv1
> > * eBGP to RIPv2
> > * eBGP to IGRP
> > * eBGP to EIGRP
> > * eBGP to OSPF
> > * eBGP to iBGP
> > * eBGP to iBGP
> > * ISIS to eBGP
>
> this is BGP routing, and you're not allowed to have two BGP processes
> running on the same router
> > * iBGP to eBGP
> > * iBGP to iBGP
>
>
> This is the final list that we need, and we also need the 'special rules'
> for classless to classful routing protocols...
> > * OSPF to ISIS
> > * IGRP to ISIS
> > * IGRP to iBGP
> > * EIGRP to RIPv1
> > * EIGRP to RIPv2
> > * EIGRP to ISIS
> > * EIGRP to iBGP
> > * RIPv1 to ISIS
> > * RIPv1 to iBGP
> > * RIPv2 to ISIS
> > * RIPv2 to iBGP
> > * iBGP to RIPv1
> > * iBGP to RIPv2
> > * iBGP to IGRP
> > * iBGP to EIGRP
> > * iBGP to OSPF
> > * ISIS to RIPv1
> > * ISIS to RIPv2
> > * ISIS to IGRP
> > * ISIS to EIGRP
> > * ISIS to OSPF
> > * ISIS to ISIS
> > * ISIS to iBGP
>
>
>
This archive was generated by hypermail 2.1.4 : Fri Jun 21 2002 - 06:45:01 GMT-3