From: Peng Zheng (zpnist@yahoo.com)
Date: Fri Jan 17 2003 - 02:25:51 GMT-3
Thanks.
--- Robert Miller <rmiller@absitech.com> wrote:
> Circumstances will determine whether or not you
> would want to use this
> method. In a real world situation, as stated in the
> book, if you are a stub
> AS with a multihomed connection to an ISP, you may
> not want to redistribute
> BGP into IGP because you don't want your routing
> tables being filled up by
> potentially hundreds or thousands of routes.
> Doyle's example is one way to
> prevent that.
>
> However, if you're in the lab and its stated that
> you are not allowed to use
> default or static routes, than regular rules of
> sychronization would
> probably apply.
>
> Later....Robert
>
>
>
> -----Original Message-----
> From: Peng Zheng [mailto:zpnist@yahoo.com]
> Sent: Thursday, January 16, 2003 7:20 PM
> To: Robert Miller
> Cc: ccielab@groupstudy.com
> Subject: RE: BGP synchronization
>
>
> Is only a default route enough? When
> synchronization
> is enabled, I think for each route to be inserted
> into
> routing table, a same route should be learned from
> IGP. Now only a default route is learned, I think
> if
> a IBGP peer learns a default route, it will put it
> into routing table. How about other routes?
>
>
>
> Thanks for help.
>
>
> --- Robert Miller <rmiller@absitech.com> wrote:
> > Follow the rules:
> >
> > 1. (pg 123) In order for IBGP to work, Peers must
> be
> > fully meshed or routes must be redistributed into
> > and syncronized with IGP.
> >
> > 2. If IBGP peers are fully meshed, syncronization
> > must be disabled in order to inject routes learned
> > from EBGP into the routing table as it traverses
> the
> > IBGP group.
> >
> > On pg 245, we know that sycronization was not
> > disabled, so IBGP could NOT inject the routes into
> > the routing table from the BGP table.
> >
> > Based on LOCAL_PREF settings, both Zermatt and
> > Moritz know from the BGP table which path they
> will
> > take to the respective AS.
> >
> > Now for the fun part!
> >
> > 1. We know that BGP did not inject the routes into
> > the routing table because syncronization was not
> > disabled.
> >
> > 2. We know that BGP is not redistibuted into
> IS-IS,
> > so this did not inject the routes into the routing
> > table.
> >
> > 3. But, Because default routes were created and
> > injected into the routing table, the IBGP peers
> are
> > able to get to their preferred paths.
> >
> > Hope this helps....
> >
> > Robert
> >
> > ------------------------------------------------
> > On Thu, 16 Jan 2003 07:53:03 -0800 (PST), Peng
> Zheng
> > <zpnist@yahoo.com> wrote:
> >
> > > I just took a look, but I didn't see the reason.
> > It
> > > only mentioned it is different with the earlier
> > > examples. I don't know why synchronization is
> not
> > > disabled.
> > >
> > >
> > >
> > > --- Robert Miller <rmiller@absitech.com> wrote:
> > > > Peng,
> > > >
> > > > Look at page 245. It explains why the use of
> > the
> > > > default route bypasses the affect caused by
> the
> > IGP
> > > > (IS-IS) and synchronization.
> > > >
> > > > Later...
> > > >
> > > > Robert
> > > >
> > > >
> > > >
> ------------------------------------------------
> > > > On Thu, 16 Jan 2003 02:24:14 -0500, "cebuano"
> > > > <cebu2ccie@cox.net> wrote:
> > > >
> > > > > Peng,
> > > > > I haven't mocked up this lab yet, but based
> on
> > the
> > > > configuration,
> > > > > Zermatt and Moritz are learning about AS50,
> 75
> > and
> > > > 100 via their eBGP
> > > > > peers as well so BGP synchronization should
> > have
> > > > not effect on these
> > > > > routers' path selection. I'll let you know
> > after
> > > > I've mocked this up.
> > > > >
> > > > > Gotta sign off now.
> > > > > Elmer
> > > > >
> > > > > -----Original Message-----
> > > > > From: Peng Zheng [mailto:zpnist@yahoo.com]
> > > > > Sent: Thursday, January 16, 2003 2:07 AM
> > > > > To: cebuano; ccielab@groupstudy.com
> > > > > Subject: RE: BGP synchronization
> > > > >
> > > > > I know the path is prefered. But because
> > > > > synchronization are not disabled on Zermatt
> > and
> > > > > Moritz, the route learned only from IBGP
> > should
> > > > not be
> > > > > put in routing table if it is not learned by
> > IGP.
> > > > >
> > > > > Right?
> > > > >
> > > > >
> > > > >
> > > > > --- cebuano <cebu2ccie@cox.net> wrote:
> > > > > > Peng,
> > > > > > Check the LOCAL_PREF values for both iBGP
> > > > routers.
> > > > > >
> > > > > > -----Original Message-----
> > > > > > From: nobody@groupstudy.com
> > > > > > [mailto:nobody@groupstudy.com] On Behalf
> Of
> > > > > > Peng Zheng
> > > > > > Sent: Wednesday, January 15, 2003 11:56 PM
> > > > > > To: ccielab@groupstudy.com
> > > > > > Subject: BGP synchronization
> > > > > >
> > > > > > Hi,
> > > > > >
> > > > > > On page 242-244 of Routing TCPIP V2, why
> > Zermatt
> > > > use
> > > > > > Moritz to 172.18.0.0?
> > > > > >
> > > > > > I think it's not synchronized because it's
> > not
> > > > > > learned
> > > > > > by IGP.
> > > > > >
> > > > > > Any idea?
> > > > > >
> > > > > > Thanks for help.
> > > > > >
> > > > > > Best Wishes,
> > > > > > Peng Zheng
> > > > > >
> > > > > >
> > > >
> > __________________________________________________
> > > > > > Do you Yahoo!?
> > > > > > Yahoo! Mail Plus - Powerful. Affordable.
> > Sign up
> > > > > > now.
> > > > > > http://mailplus.yahoo.com
>
=== message truncated ===
This archive was generated by hypermail 2.1.4 : Sat Feb 01 2003 - 07:33:52 GMT-3