Page 205

OSPF Scalability

167

OSPF Scalability

n the previous chapter, we examined the configuration of OSPF net-

works that contained a single area. We saw that OSPF had significant advan-
tages over distance-vector protocols, such as RIP, due to OSPF's ability to
represent an entire network within its link state database, thus vastly reduc-
ing the time required for convergence.

However, let's consider what the router does in order to give us such great

performance. Each router recalculates its database every time there is a
topology change, requiring CPU overhead. Each router has to hold the entire
link state database, which represents the topology of the entire network,
requiring memory overhead. Furthermore, each router contains a complete
copy of the routing table, requiring more memory overhead. Keep in mind
that the number of entries in the routing table may be significantly greater
than the number of networks in the routing table because we may have mul-
tiple routes to multiple networks.

With these OSPF behavioral characteristics in mind, it becomes obvious

that in very large networks, single area OSPF has some serious scalability
considerations. Fortunately, OSPF gives us the ability to take a large OSPF
topology and break it down into multiple, more manageable areas, as illus-
trated in Figure 5.1.

F I G U R E 5 . 1

OSPF areas

Single area OSPF network

Multi-area OSPF network

Area 0

Area 10

Area 20

www.sybex.com