Section 7.2. Area Reliability

7.2. Area Reliability

A reliable area is one that can sustain the failure of one or more of its components without isolating a large segment of the network. The more failures the area can tolerate, the more robust the area.

Large network segments can be isolated in one of two ways:

• The failure of a router connecting two areas isolates an area from the rest of the network (Figure 7.2).

  Figure 7.2. A failure of one of the ABRs will isolate its attached nonbackbone area from the rest of the network.

  

  
  

• The failure of a router or link partitions an area (Figure 7.3).

  Figure 7.3. A failure of any one of the three links connecting the ABRs across area 0 will partition area 0, resulting in the isolation of one or more areas from the rest of the network.

  

  
  

The obvious way to avoid the isolation of an area from a failed ABR is to provide multiple ABRs and multiple paths to each ABR, as shown in Figure 7.4. In this network, area 2 cannot be isolated from area 0 by the failure of a single ABR or ABR interface. Likewise, no router in the area can be isolated from the rest of the network by the failure of a single, or even any two, links or interfaces.

Figure 7.4. A robustly connected nonbackbone area.

Similarly, robust connectivity in the backbone area depicted in area 2 of Figure 7.4 ensures that the backbone cannot be easily partitioned. This area is important enough that at the least it should be designed so that no two link or router failures can partition the area. Additionally, every router connecting a nonbackbone area to the backbone should have connections to at least two other routers in the backbone. This further reduces the chance of isolating a nonbackbone area.

Just because you can divide your network into areas does not mean you must. Many networks can avoid the potential pitfalls of poor multi-area designs by simply using a single area for their entire IGP domain. When designing your network, consider all the factors influencing the scale of the network, the primary purpose of the network and its IGP, and the future growth of the network. Then consider what advantage multiple areas provide, and weigh them against the potential advantages of a single area. As you will see in the remainder of this chapter and the next chapter, such factors can influence not only how you design your network, but also what IGP you choose.