RE: EBGP Multihop's necessity with loopback addresses

From: Nathan Chessin (nchessin@cisco.com)
Date: Wed Oct 16 2002 - 19:35:01 GMT-3


I think your view of the ebgp-multihop command is incorrect. It is called
ebgp-multihop for a reason. Not ibgp-multihop. The update-source is used
for loopbacks to source the BGP info, this is correct. But not because you
are more than one hop away. For example, you could have a neighbor x.x.x.x
command where x.x.x.x is on the other side of the router from where you are
physically connected. BGP uses TCP connections to distinguish neighbors.
Also, if you have all of them peering in an AS you have to have neighbor
statements to all of them, unless using RR or Confeds, but I digress.

http://www.cisco.com/warp/customer/459/13.html#A5.0

You only need the ebgp-multihop command for eBGP.

Nate

> -----Original Message-----
> From: nobody@groupstudy.com [mailto:nobody@groupstudy.com]On Behalf Of
> Paglia, John (USPC.PCT.Hopewell)
> Sent: Wednesday, October 16, 2002 2:32 PM
> To: 'Joe Martin'; ccielab@groupstudy.com
> Subject: RE: EBGP Multihop's necessity with loopback addresses
>
>
> I think the whole basis of this thought has to do with your
> next-to-last
> sentence.
>
> The 'update-source' command (and correct me if
> mistaken...like y'all need to
> be told that!!!!) tells the neighbor router specified in the
> command to
> observe YOUR loopback as the source of the peering
> relationship. Thus, the
> interface that is being utilized for peering, from the point
> of view of the
> neighbor, is not directly connected, but is at least 2 hops
> away, whether it
> be EBGP or IBGP.
>
> I do really think this is outdated but I just thought I'd
> throw it out there
> and see if anyone knew for sure, and judging from the replies
> thus far, it
> has to be an 'out-of-practice' practice.
>
> Thanks,
> John
>
> > -----Original Message-----
> > From: Joe Martin [SMTP:jmartin@capitalpremium.net]
> > Sent: Wednesday, October 16, 2002 4:58 PM
> > To: ccielab@groupstudy.com
> > Subject: RE: EBGP Multihop's necessity with loopback addresses
> >
> > If you use a local interface or a physical interface ip
> address in your
> > bgp
> > neighbor statement then you would lose your peering if the
> interface or
> > network goes down, even if there is another route to the
> bgp neighbor.
> > this
> > is why it is suggested that you use the loopback ip as your
> update source
> > and specify the loopback address in your neighbor
> statements. This way,
> > if
> > your direct connection to your bgp neighbor goes down you can still
> > maintain
> > the neighbor relationship. However, now your neighbor is
> not neccessarily
> > 1
> > hop way. So this is where the ebgp multi-hop command comes in.
> >
> > Someone correct me if i'm wrong.
> >
> > -----Original Message-----
> > From: nobody@groupstudy.com
[mailto:nobody@groupstudy.com]On Behalf Of
> Paglia, John (USPC.PCT.Hopewell)
> Sent: Wednesday, October 16, 2002 2:01 PM
> To: 'ccielab@groupstudy.com'
> Subject: EBGP Multihop's necessity with loopback addresses
>
>
> I recently heard that if you are establishing your BGP neighbors using
> 'update source loopback 0', you should also use the 'ebgp-mu' cmd, even if
> the neighbors are directly connected...the reason being that your loopback
> is NOT directly connected to the neighbor. However, in my experiments I
> have
> never done this for neighbors that are directly connected, yet have
> established peerings successfully.
>
> Is there validity to this statement, and if so, under which circumstances
> is
> it absolutely vital, other than the 'non-physically or nbma topology'
> scenarios??? Something tells me that this may be an older IOS issue or
> something like that.
>
> John



This archive was generated by hypermail 2.1.4 : Tue Nov 05 2002 - 08:35:48 GMT-3