There are other paths through the
network, but their hop counts exceed both the primary route and the
feasible successor. It is possible, for example, for router C to forward
datagrams to network 193.9.1 by using the route through router B to D to E
and, finally, to router A and network 193.9.1. However, the hop count of
this network (where both bandwidth and delay are equal across all links)
is four. Therefore, it is unattractive as both a primary route and a
feasible successor to the primary route. Such a route may become either a
primary route or a feasible successor, but only if multiple network
failures occurred and it were the least-cost route. However, the entire
EIGRP network would have to recompute routes to known destinations for
this to occur. To understand how the process of finding an alternative
path works, consider Figure . In this illustration, the link
between routers B and C fails.
The consequences of this failure are that
all routes that used B as a next hop go active in the EIGRP topology
table. The effects of this failure, as documented in the topology table,
are summarized in Figure .
In this
example, all routes that used the link between routers C and B become
active in the topology table. Other routes, including those that pass
through router B via router A, remain passive and unaffected by the
topology change.
Router C responds to this topology change by sending a
query to its neighbors, notifying them that it has lost two primaries. It
has only two neighbors, B and A, and one of them is now unreachable.
Router A is obligated by the protocol specifications to
respond to the router C query for alternative path information. Its own
topology table has not been affected by the link failure because it has a
different set of neighbors. Therefore, there is hope that other routes can
be discovered.
The router A topology table, in the middle
of convergence, is summarized in Figure .
The link failure between routers B and C has not
affected any of the router A primary routes. They remain passive and in
use. Therefore, router A will respond with information on an alternative
route through router E to these destination networks.
When router C receives the reply from router A, it knows
that all the neighbors in the network have processed the link failure and
modified their tables accordingly.
Figure
summarizes the results of the new understanding that router C has of the
network topology.
Router C was able to
identify an alternative path --- that is, successor --- to all the routes it
had been able to reach through router B. These alternatives are far from
ideal, however: They all begin with the hop to router A, which is also the
primary route to 193.9.1. If a failure were to occur on this link, router A,
or on any of the router interfaces that connect this link to routers A and C,
router C would be completely isolated from the rest of the network.
|