Advertsing Loopback in BGP and EIGRP

FrankGuthrie

I was working n the INE workbook for CCIE R&S and in on of the tasks (BGP Peer Groups) I see that in both BGP and EIGRP the loobacks are advertised. See the attached diagram:


In the config on R1 -R8 the loopbacks (150.1.x.x - where X is the Routers number) are advertised in both EIGRP and BGP.


If I look at the Routers in AS100, for example R3:
R3#sh ip bgp
BGP table version is 29, local router ID is 150.1.3.3
Status codes: s suppressed, d damped, h history, * valid, > best, i - internal,
r RIB-failure, S Stale, m multipath, b backup-path, f RT-Filter,
x best-external, a additional-path, c RIB-compressed,
Origin codes: i - IGP, e - EGP, ? - incomplete
RPKI validation codes: V valid, I invalid, N Not found




Network Next Hop Metric LocPrf Weight Path
*>i 28.119.16.0/24 150.1.7.7 0 100 0 54 i
*>i 28.119.17.0/24 150.1.7.7 0 100 0 54 i
*>i 112.0.0.0 150.1.7.7 0 100 0 54 50 60 i
*>i 113.0.0.0 150.1.7.7 0 100 0 54 50 60 i
*>i 114.0.0.0 150.1.7.7 0 100 0 54 i
*>i 115.0.0.0 150.1.7.7 0 100 0 54 i
*>i 116.0.0.0 150.1.7.7 0 100 0 54 i
*>i 117.0.0.0 150.1.7.7 0 100 0 54 i
*>i 118.0.0.0 150.1.7.7 0 100 0 54 i
*>i 119.0.0.0 150.1.7.7 0 100 0 54 i
r>i 150.1.1.1/32 150.1.1.1 0 100 0 i
r>i 150.1.2.2/32 150.1.2.2 0 100 0 i
*> 150.1.3.3/32 0.0.0.0 0 32768 i
r>i 150.1.4.4/32 150.1.4.4 0 100 0 i
r>i 150.1.5.5/32 150.1.5.5 0 100 0 i
r>i 150.1.6.6/32 150.1.6.6 0 100 0 i
r>i 150.1.7.7/32 150.1.7.7 0 100 0 i
r>i 150.1.8.8/32 150.1.8.8 0 100 0 i


You can see that the loopbacks gotten from BGP are not put in the Global Routing table. This has to do (I think) with the fact the Routers in AS 100 get the loopback also from EIGRP, as EIGRP has an AD of 90 and BGP of 200.


My question was why are they advertised both in EIGRP and BGP?


I think I have the answer and need some one to verify is this is correct. So the reason to advertise it in BGP is for R9 and R10 to get the loopback as they're are not in the same EIGRP AS. The Routers in EGRP AS 100, R1 - R8, will use the EIGRP routes, correct? But if the Router in AS 100 are speaking BGP, why do we need to advertise the loopbacks in EIGRP?

_Gonzalo_

FrankGuthrie wrote: »

You can see that the loopbacks gotten from BGP are not put in the Global Routing table. This has to do (I think) with the fact the Routers in AS 100 get the loopback also from EIGRP, as EIGRP has an AD of 90 and BGP of 200.

I would say that you are right here. When BGP table´s routes have an "r" instead of an "*", it usually means that.

FrankGuthrie wrote: »

My question was why are they advertised both in EIGRP and BGP?


I think I have the answer and need some one to verify is this is correct. So the reason to advertise it in BGP is for R9 and R10 to get the loopback as they're are not in the same EIGRP AS. The Routers in EGRP AS 100, R1 - R8, will use the EIGRP routes, correct? But if the Router in AS 100 are speaking BGP, why do we need to advertise the loopbacks in EIGRP?

This is something else... The way I understand it, advertising them on both protocols makes no sense. You want to have your IGP (EIGRP in this case) advertising the loopbacks of your AS 100, so your inside peering relations are simpler to configure (especially for a full mesh)
Even if you wanted those nodes in other AS were able to ping these loopbacks, you would redistribute from EIGRP to BGP or just use BGP for that.

Regarding the other AS, you will not have an IGP here, as it would not make any sense (that´s what BGP is for!) The only peering options I know of are:

1- Peer with the directly connected IP (no need to advertise anything here)
2- Peer with loopback. Static routes are usually configured for this purpose, as it would make no sense that they would speak any IGP between two peers from two different AS, and obviously BGP is not an option here...

So it makes sense to advertise loopbacks in AS 100 using EIGRP, and (if there is a point for it) advertise loopbacks on borderline routers using BGP.

rjon17469

FrankGuthrie wrote: »

You can see that the loopbacks gotten from BGP are not put in the Global Routing table. This has to do (I think) with the fact the Routers in AS 100 get the loopback also from EIGRP, as EIGRP has an AD of 90 and BGP of 200.


My question was why are they advertised both in EIGRP and BGP?

This is an important aspect of BGP and the different ADs between eBGP and iBGP routes. iBGP is purposefully designed to have RIB-failures for internal routes, because BGP relies on an IGP for internal routing. Sure, you can use BGP in place of an IGP if you really want to, but it won't be pretty.

The loopbacks are advertised into both for a couple reasons. First, they go into the IGP for internal connectivity, as that's the IGP's function. Then, they go into BGP to support synchronization (if you're a glutton for punishment) as well as for BGP multihop, which is by default enabled for iBGP peers. And finally, once they are in BGP they can be advertised to neighbors, as you mentioned.

FrankGuthrie

Hi Guys,

I think I have an explanation.

So the loopback of all the routers are advertised using EIGRP within AS 100, but for R9 and R10 which are in an other AS, BGP is used. The Routers in AS 100 will get a RIB failure, but still receive the route from EIGRP. That way connectivity between the loopbacks in the AS and outside of the AS is propagated.

_Gonzalo_

As for what I posted, that should be it. Just a small correction:

FrankGuthrie wrote: »

So the loopback of all the routers are advertised using EIGRP within AS 100, but for R9 and R10 which are in an other AS, BGP is used. The Routers in AS 100 will get a RIB failure, because they receive the route from EIGRP. That way connectivity between the loopbacks in the AS and outside of the AS is propagated.

And this rib-failure will happen if loopbacks are advertised in both EIGRP and BGP.

What I did not catch is what rjon17469 said about the advantage (if it wasn´t irony) of having those loopbacks advertised in BGP as well:

rjon17469 wrote: »

Then, they go into BGP to support synchronization (if you're a glutton for punishment) as well as for BGP multihop, which is by default enabled for iBGP peers. And finally, once they are in BGP they can be advertised to neighbors, as you mentioned.