I've actually heard that term used before, and I think it *may* have
been in one of Howard's books or CertificationZone papers. Maybe not,
I'm sure he'll correct me. I do know that I've heard the term "split
horizon" in this context before.
I agree that we should probably use it more often. A slight
modification might be that a BGP speaker will ignore updates that
include its own ASN in the path. I don't know if that still qualifies
as split horizon since it's the receiver of the route that is
disallowing the update, not the sender (as in the case of split horizon
in other protocols.) Still, the end result is the same and I think it
applies.
John
>>> "Chuck Larrieu" 8/21/01 9:29:20 AM >>>
As a result of an off-list conversation, I came across the following
networking term:
iBGP split horizon
my first reaction was a sarcastic remark about never having come across
the
term in the RFC's. but then I got to thinking about it, and I now see
this
as a descriptive and quite useful term.
recall that distance vector protocols are subject to the rule of split
horizon. they do not advertise a route back out the interface that
they
received that particular route.
one of the gotchas of iBGP is the fact that iBGP routers do not
propagate
routes learned from one iBGP neighbor to other iBGP neighbors. hence
the
requirement for iBGP full mesh.
so why not call this iBGP split horizon? and define it as follows: an
iBGP
router will not advertise a route back out the same AS from which it
learned
the route?
does this make sense? worth letting this one join the lexicon of
networking
terminology?
Chuck
Message Posted at:
http://www.groupstudy.com/form/read.php?f=7&i=16686&t=16668
--------------------------------------------------
FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html
Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
