From: Soup Shi (soup_ccie@hotmail.com)
Date: Mon Dec 08 2003 - 02:12:03 GMT-3
Hi, group,
I know it is an old topic and must be discussed over and over.
However, I can not figure it out even I feel I understand it completely.
Say a router learns a route from both IGP, EBGP and IBGP, the IBGP has a
shorter as-path than EBGP. From my understanding, the IGP route would get
into routing table finally, because IBGP and EBGP routes got compared first
in the BGP table and the IBGP wins, then it failed when compared with IGP
route when going towards routing table since its bigger AD.
However, when the synchronization is on. There is no guarantee at all! It
depends on who comes first. If the EBGP route comes earlier than IGP, it
gets into routing table first, so the IBGP is never valid in BGP table. The
result is IGP loses its competition since its bigger AD.
Is there any ways to avoid this lottery picking? I think most of the time in
a lab, the only reason for redistribution between BGP and IGP is that we are
not allowed to turn off the synchronization. So do we have to rely on pure
luck in a CCIE lab exam?
Thanks in advance.
Soup
This archive was generated by hypermail 2.1.4 : Sat Jan 03 2004 - 08:25:37 GMT-3