A device (computer or router) may have no static route or dynamically learned route to a particular destination within the network. A static route may be absent because the network administrator never defined one, or because the software does not support defining one (e.g., end stations often do not support static routes). Learned routes might be absent if the device has no ability to learn them (also commonly true of end stations), if it has not had time to learn one, or if the routing protocol explicitly bars it from learning one. If there is no static route or learned route, the Network Layer checks for a default route.
The Role of the Default Route
Default routes serve an important function. Consider the Internet and all of its subnetworks. Maintaining a list of all possible destinations is well beyond the ability of most routers, and certainly beyond the ability of the end stations. Imagine the memory required to store thousands of IP addresses, each four octets long, together with the route information for how to reach that network (not to mention the processor time and bandwidth required to run the protocol that would keep them all up-to-date)! And if the company has a single link to the Internet, all of that information is largely useless because all of those networks can be reached by passing the packets to the same next-hop router: the one owned by the ISP providing Internet access. It makes far more sense for this router to have detailed information about all of the networks within the corporate network, so it can forward inbound packets correctly, but a single default route (for outbound traffic) representing all networks out in the Internet.
Default Route in IP
Although the concept of the default route can occur in any packet network, most modern packet networks are based on IP, so our discussion will focus on that protocol.
In IP, the default route is also known as the default gateway or the gateway of last resort. The term gateway is an old one and reflects the roots of the IP protocol. In the earliest days of internetworking, the device that interconnected the various networks in the Internet were known as gateways. However, that term came ot have a new meaning due to its use by the ISO and the term router became more common.
The default route is represented using a special IP address (0.0.0.0) and subnet mask (0.0.0.0). This address is placed in the routing table along with the address of the local router to be used (i.e., the default gateway). In IP, routers and end stations use a simple rule when selecting a route from the routing table: the best match wins. So if a routing table contained entries for 184.108.40.206, 220.127.116.11, 192.0.0.0, and 0.0.0.0, and a packet arrived destined for 18.104.22.168, the route associated with 22.214.171.124 would be used because it has the greatest match with the destination. If a packet arrived destined for 126.96.36.199, however, the route associated with 188.8.131.52 would be used. By extension, if a packet arrived destined for 184.108.40.206, the default route of 0.0.0.0 would be used. This route matches every IP address not matched by any other entry in the routing table, making it the default.
Implementing the Default Route
As with any entry in the IP routing table, the default route can be placed there statically, or learned dynamically. It is typically entered statically by a network administrator in a router that connects by a single path to the broader internetwork. For example, if there was a large regional branch location that had a single connection back to the data center and the rest of the corporate network, the router at the branch could be configured with a static default route pointing back to the router in the data center. A router connected to an ISP could be configured with a static default route pointing to the ISP's router.
If there are multiple subnetworks behind the router with the statically configured default route, these other routers could learn about the default route using the routing protocol. In this case, the router with a statically configured default route in known as the originating router. In our example of the router connecting the regional branch to the data center, this router could teach any routers within the regional branch that it is the "go to" router if they receive packets bound for locations outside the regional branch. The router connecting a corporation to the Internet could teach the routers within the corporation that it is the "go to" router for any traffic intended for networks outside the corporation (e.g., out in the Internet).
When originating the default route, it is important to ensure that the originating router be as close as possible to the exit point of the network. Otherwise, routing loops can result.
|<mp3>http://podcast.hill-vt.com/podsnacks/2008q3/default_route.mp3%7Cdownload</mp3> | Default route|