HP ProCurve Secure Router 7203 dl Advanced Management And Configuration Manual page 630

IP Routing—Configuring RIP, OSPF, BGP, and PBR
Configuring RIP
13-16
As long as the network remains stable, this process continues smoothly.
However, problems arise if the topology changes.
Consider what happens when the link between Router B and Network 1 fails.
(See Figure 13-2.) Router B begins advertising a route to Network 1 with a
metric of 16 to indicate that it is unreachable. Routers A and C receive this
update from Router B and change the metric, but not before they have already
sent their own routes for Network 1 with a metric of 2. Router A receives the
route from Router C, and Router C receives the same route from Router A.
Because these routes have a lower metric than the route through Router B,
Routers A and C store these routes in their routing tables (adding one to the
metric). (See Figure 13-2.) Router B may receive the route from either Router
A or C. As mentioned earlier, Router B has no way to determine that this route
ultimately points back to itself and so is invalid. Because its own connection
to Network 1 has failed, Router B accepts the route.
Router A
Before connection fails
Destination Next Hop Metric
* 1 B
1 A
1 D
Figure 13-2. Count to Infinity
Routers A and C now each have a route to Network 1 with a metric of 3, pointing
to each other. In the next update cycle, Router A receives the route from Router
C. It updates the route in its table with a metric of 4. Router C, receiving the
update from Router A, does the same. The next time the routers advertise the
route, it has a metric of 4. Eventually, the metric will reach 16, and the routers
will determine that they cannot reach Network 1 through each other. This
process is called a count to infinity, and it can significantly slow convergence.
Network 1
Router B
Router C
After connection fails
Destination Next Hop Metric
2 1
3 * 1
5 1
4 hops
Router D
B 16
A 3
D 5