Highly Available Networks

Figure 10 shows you how to isolate the clients on separate subnets via routers. Like the previous
example, the local subnet for the servers is made highly available by using two cable segments and a
hub. However, connection to the servers is made via the routers. The clients have a single point of
failure at the router. If a router fails, the clients connected via that router will be unable to communicate
with the server. The client’s connections through the other router will still be up and communicating. By
using routers, the clients have been isolated just as with the HUBs. A single router failure will still cause
half of the clients to be disconnected from the servers.
Figure 11: Highly Available Bus Topology with Cross-Connected Routers
In Figure 11, each router is connected to each LAN. The client LANs have access to the servers through
both routers. The advantage of using routers over HUBs becomes clear when we cross connect the
routers and start using gated protocols to provide current subnet status. The routers will poll each other
for current subnet access maps and broadcast this information to anyone listening. The clients and
servers need to be configured in listen only mode to receive current route information from the routers.
The clients and servers should not be using the IP broadcast address (for example, by configuring the
default route as the local system) to route traffic to other subnets. Instead they should be using specific
router IP addresses and gated-configured primary and alternate route paths. In this way the maximum
throughput of the LAN is preserved (for example, no broadcast storms are being created) and errors such