The answer is in the question...
You said you are using subinterfaces. No doubt your hub site has a
multipoint subinterface. In this case, the router treats each subinterface
as if it were a seperate physical interface. So there is no issue with
split-horizon.
The problem with split-horizon come into play with physical interfaces. If
you have multiple PVC's associated with a single physical interface, you
could have a problem. Let's say your hub has multiple PVC's (one for each
spoke) associated with a single physical interface, s0. A route that is
learned by the hub from a spoke cannot be advertised to the other spokes
because split-horizon will prevent it from being sent back out the same
interface on which it was received (s0). In this instance, you need to turn
off IP split-horizon at the hub.
Change the hub to subinterfaces, and a route learned from one spoke comes in
on one subinterface (let's say s0.1), and the router has no problem
readvertising the route on subinterfaces s0.2 and s0.3 because it treats
them as if they were seperate physical interfaces.
Is this clear? or am I rambling...
David's paper states:
"With frame-relay, a single physical interface can have many different PVCs
associated with it.
...
R1 will learn about the 10.0.0.0/8 network via its own serial 0 interface,
but split-horizon will prevent R1 from ever advertising the 10.0.0.0/8
network back out the serial 0 interface."
He specifically states that this applies to physical interfaces, whic is not
your configuration. So there is no contradiction...
I hope this helps...
Dale
[=`)
>From: "Scott Livingston" <[EMAIL PROTECTED]>
>Reply-To: "Scott Livingston" <[EMAIL PROTECTED]>
>To: [EMAIL PROTECTED]
>Subject: [Fwd: Morning]
>Date: Tue, 06 Jun 2000 12:15:55 -0500
>
>can someone help out here please? THANKS!
>subject:
> Morning
> Date:
> Tue, 06 Jun 2000 07:50:00 -0500
> From:
> Scott Livinston <[EMAIL PROTECTED]>
> To:
> [EMAIL PROTECTED], Jon Helmer <[EMAIL PROTECTED]>
>
>
>
>
>
>This is reference to Mr. David Wosefer's white paper about Frame Relay..
>
>David,
>
>We are currently running a small F.R. network over here and something
>you wrote contradicted what we currently have in operation.. Currently
>we have a hub and spoke topology, 3 spokes to be exact... we are an IP,
>partial meshed network running subinterfaces and EIGRP.. You mentioned
>in your paper that you need to turn off ip split-horizon on the
>subinterfaces if the spokes are to know about the other spokes
>networks.. Well in our case we don't have ip split horizons turned off
>and the spokes know about all other spokes... How could this be? I had
>my lead engineer look @ this and we both cant figure out this
>discrepancy between what you published and what we are currently running
>
>over here... If you get some time could you please show me where i might
>
>be lost? Thanks for your help!
>
>--------------------------------------------------
>Scott M. Livingston
>Network Engineer (CCNP)
>12851 Foster
>Overland Park, KS 66213
>800.888.7535
>913.402.7844 x1056
>913.814.7849 Fax
>
>"Make every swing as if it were your last"
> -Gary Schroer
>--------------------------------------------------
><< message3.txt >>
________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com
___________________________________
UPDATED Posting Guidelines: http://www.groupstudy.com/list/guide.html
FAQ, list archives, and subscription info: http://www.groupstudy.com
Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
