Conceptually, the simplest method for maintaining routing tables is to configure them into the routers manually. This method naturally relies on the accuracy of the network administrator to avoid problems such as loops. It is also up to the administrator to update the tables whenever something changes anywhere in the network.
Maintaining a large network using static routes presents several problems, however. It is cumbersome and labor intensive. It is impossible to achieve automatic fault tolerance because changing routes in response to a network failure always takes time. It also demands that the network administrator have perfect knowledge of the state of the network.
Furthermore, when building static routing tables on a network of any size, making mistakes that isolate remote parts of the network is remarkably easy. When this happens, technicians may need to visit the remote devices and manually reconfigure them from the console. Put simply, static routing is not an effective way of handling the main routing of a network. However, it still has its place, even in a network that uses a more sophisticated dynamic routing protocol to build its tables. For example, when connecting to external networks, it is often easier to configure a few static routes than to share routing information with the external network. This is particularly true when the external network is always accessed through one point. If this Access point goes down, the network doesn't have a backup path, so there is no point in updating the route to elsewhere.
In fact, it may be worse to try more sophisticated methods. Suppose there is more than one external network, such as an Internet connection and a separate secure connection to a partner organization. If that secure connection becomes unavailable, you probably don't want the packets sent out to the Internet instead. In all likelihood, this would happen if the network exchanged routing information with the partner organization.
This is because the partner network's IP-address range is not part of the internal range. When the dynamic routing information from that network disappears, the global default static route pointing out to the public Internet is used instead. Depending on the specific type of failure and the exact network configuration, the same thing could happen if you used a static route to the partner network. For the purposes of this example, let me assume that the static route was added in a way that allows it to remain in effect even if the link goes down.
You may need to use static routes in networks involving equipment that doesn't support the preferred dynamic routing protocols. In this case, though, you would only use the static routes to get through these isolated parts of the network.
Static routes definitely have their place, but they should be used sparingly. Each time you configure a static route you have to ask whether it would be better if this routing information were learned dynamically. As you will see in the following discussion, once a static route exists on one of your routers, it is relatively easy to use the dynamic routing protocol to distribute this information throughout the rest of the network.
Another important kind of static route is a floating static route. This feature is not available on all vendors' routers. A floating static route is like a normal static route, but it isn't always present. That is, if a better route is available, the router will not look at this static route. But if that better route disappears, then the router will revert to the floating static route.
The way it works is simple enough in concept. The floating static route is manually configured in the router, the same as any other static route. But it has an extremely high metric to indicate a high cost associated with this path. If there is any other path available, it will be better than this one and consequently will not be used.
This feature is commonly used in dial-backup situations, for example. When the network is working properly, a remote router receives its routing table via a dynamic routing protocol. However, when there is a failure, the router stops receiving any dynamic routing information, and it flushes all of this dynamic information out of its routing table. When that happens, the floating static route suddenly starts to look good despite its high metric. The router inserts this route into its routing table, and this triggers the dial-backup process.
The key to floating static routes is that there is a magic metric value. Normally, if a router has a static route, it will use it. On Cisco routers it is conventional to use a metric of 200 or greater for floating static routes, although values as low as 150 appear to work just as well.