Information Technology Reference
In-Depth Information
next-hop-self
. When R1 reflects the route to R3 and R4, suppose it resets the next hop to
itself. R4 forwards the traffic to the destination of 172.16.0.0/16 to R1. Because R1 and R2
are not physically connected (an improper design, as indicated before), R1 forwards the
packet to R2 via R4. This results in a routing loop.
Figure 7-28
Improper Next-Hop Setting Leads to Routing Loops
AS 100
AS 200
NEXT_HOP:
192.168.1.1
192.168.1.1
192.168.1.2
RR
R1
R5
R2
NEXT_HOP:
192.168.1.2
172.16.0.0/16
NEXT_HOP:
192.168.1.2
Client
Client
R3
R4
Physical
iBGP
The topology shown in Figure 7-28 uses an RR design that does not follow the basic RR
design guidelines that were discussed previously. The topology is used here to demonstrate
the importance of proper next-hop settings.
NOTE
Under certain conditions, however, it might be desirable to reset the BGP next hop on RRs.
One example is to bring RRs into the forwarding path. IOS provides methods to accomplish
this. The following discusses how this can be done.
Basically, you can use two methods to set the BGP next hop at an RR:
•
neighbor next-hop-self
command
•
Outbound route map
If the routes are learned from an external peer, you can reset next hops on RRs by using
next-hop-self
or an outbound route map to clients and nonclients. Note that RRs here are
border routers. Figure 7-29 shows an example.
