Unlike most LAN components, routers are intelligent. More importantly, they can operate at all layers of the OSI reference model rather than just the first two. This enables them to internetwork multiple LANs by using Layer 3 addressing.

A router must have two or more physical interfaces for interconnecting LANs and/or WAN transmission facilities. The router learns about the addresses of machines or networks that are somehow connected via each of its interfaces. The list of these addresses is kept in tables that correlate Layer 3 addresses with the port numbers that they are directly or indirectly connected to.

A router uses two types of networking protocols, both of which operate at Layer 3. These are routable protocols and routing protocols. Routable protocols, also known as routed protocols, are those that encapsulate user information and data into packets. An example of a routed protocol is IP. IP is responsible for encapsulating application data for transport through a network to the appropriate destinations. Routing protocols are used between routers to determine available routes, communicate what is known about available routes, and forward routed protocol packets along those routes. The purpose of a routing protocol is to provide the router with all the information it needs about the network to route datagrams.


Routers are used to forward packets of data between devices that aren't necessarily connected to the same local network. Routing is the cumulative processes that discover paths through the network to specific destinations, compare redundant routes mathematically, and build tables that contain routing information.

In the sample internetwork presented in Figure 1-8, the router's task is easy: It has only two interfaces.

An Introduction to Internetworking

Any packets received by one of its interfaces was either delivered to the other interface or discarded as undeliverable. In this particular case, the router may well have been replaced by a hub, bridge, switch, or any other Layer 2 device. The router's real value lies in determining routes to destinations on nonadjacent networks. Figure 1-10 illustrates this scenario.

Figure 1-10: Routers can calculate routes through other networks.

Figure 1-10: Routers can calculate routes through other networks.

In such circumstances, where the destination is not directly reachable, the router can communicate with other routers to learn of the existence of any networks that are not directly connected to it. In the example presented in Figure 1-10, User 1 on a network connected to Router A must access files stored on Server 2. Unfortunately, that server is located on a network that is not directly connected to Network A. This server is two router hops away on the network connected to Router C.

The way that Router A can figure out where to send User 1's packets is rather simple. Upon receiving User 1's data frames, it unwraps the packets that they contain and reads the destination IP address contained in the packet headers. Router A looks up that address in its routing table. The routing table entry tells the router that, although that server is not directly connected to the router, the next step in the path to that server is its S0 (for Serial Port 0) interface.

The S0 interface connects Router A with Router B. Router B accepts the incoming packets (albeit in a stream of bits) from Router A and looks up the destination IP address in its routing table. Router B's table identifies its S1 interface port as the next step, or hop, in the route to Server 2.

Router B forwards all packets addressed to Server 2 through its S1 port, which connects to Router C. Router C can identify the host as residing on its directly connected FDDI network and, therefore, wraps the packets in FDDI frames. These frames bear the MAC address of the router's FDDI port as a source address and the MAC address of Server 2 as the destination address. After the framed data arrives at Server 2, that server's protocols are responsible for checking the validity of the data and forwarding it up through the various layers to the correct application.

This highly simplified overview of a routed connection betrays the complexity of routing and, more importantly, calculating routes through complex networks (which is discussed in the next section). In complex networks, there may be multiple potential paths through a network between any given pair of source and destination machines. Routers are responsible for sharing information among themselves.

An Introduction to Internetworking

This information enables them to

• Discover and track the topology of the internetwork

• Discover and track the addresses of subnetworks and hosts

• Discriminate between optimal and suboptimal routes

• Balance loads across redundant optimal routes

These capabilities are examined in much more detail throughout this book. Calculating Routes

Routers can discriminate among multiple potential paths to select the best one. This process is known as calculating routes. Implicit in this description is that there is some logic and mathematics that can be applied in determining routes to far-away destinations. The capability to apply this logic and perform these mathematics is the single most important feature of a router.

The technology that enables routers to calculate routes is known as a routing protocol. Actually, there are a wide variety of routing protocols to choose from. Most are widely supported enough to permit the construction of internetworks from routers made by different manufacturers.

Routing protocols enable routers within a network to share information about potential paths to specific hosts within that network. Examples of routing protocols include the following:

• Routing Information Protocol (RIP)

• Open Shortest Path First (OSPF)

• Interior Gateway Routing Protocol (IGRP)

These, and several other routing protocols, are examined in more detail in Part III, "Routing Protocols." Each routing protocol has its own unique blend of features, benefits, and limitations. Many of them are also well targeted to specific functional niches. Consequently, in very large networks such as the Internet or even some large corporate intranets, it is quite likely that you will encounter two or more routing protocols in operation. Therefore, it is a good idea to become as familiar as possible with as many of these protocols as you can. "Internetworking with Dissimilar Protocols," describes some of the ways that you can get two dissimilar routing protocols to peacefully coexist within the same internetwork.

0 0

Post a comment