Network Warrior

Chapter 13. Resilient Ethernet

When designing a network, eliminating single points of failure should be a priority for any network engineer or architect. While it may be easy to assume that having two of every device will provide redundancy, how does one go about truly making the devices redundant?

Devices like PIX firewalls and CSM load balancers have redundancy and fault-tolerance features built into their operating systems, which even go so far as to transfer configuration changes from the primary to the secondary devices. Cisco routers don't really have that level of functionality, though, and with good reason. While you may wish to have two routers be a failover default gateway for a LAN, those two routers may have different serial links connected to them, or perhaps a link from one Internet provider connects to one router, while a link from a different provider connects to the other. The router configurations will not be the same, so configuration sync will not be practical.

Usually, on routers we're looking for the ability for one device to take over for another device on a specific network. Routers generally support multiple protocols, and connect many types of technologies, and each technology can be configured with the failover method preferred for that technology. In the case of Ethernet, the methods most often used are the Hot Standby Router Protocol (HSRP) and the Virtual Router Redundancy Protocol (VRRP). HSRP is Cisco-specific, while VRRP is nonproprietary, and thus ...