Improving Network AvailabilityUpdated 7/93For companies that rely heavily on network applications, a failure in thenetwork can be disastrous. Network failures are more likely to occur asnetworks grow—the natural result of employing more and more networkequipment. This application note examines some of the methods forincreasing network availability (uptime) in router-based networks andthereby reducing or eliminating user-perceived network failures.The most common technique for increasing network availability is toprovide alternate paths for data in the event of link or router failures. Thisplays to the strength of routers—quickly detecting network failures androuting data around them. This application note examines the different waysof providing alternate data paths to survive both data-link and router failures.Permanent Alternate PathsConsider the network in figure 1. There is only one path for data from any siteto any other site. Should link 1 fail, for instance, systems at site A will be unableto communicate with systems at either site B or site C, until link 1 has beenrestored. Adding a third link, as indicated by the dotted line between the siteA and site C routers, provides an alternate path for data to and from sites A andC. Once the third link has been installed, any single link failure will not resultin a user-perceived network failure.Application Notes/Case Studies33-3