Moreover, RIP (version 1) wastes valuable network bandwidth by propagating routing information via broadcasts, nor does it consider enough network variables to make the best routing decision.

4-8

6.2.6.2 OSPFv2 Dynamic Routing Protocol

OSPF overcomes all the problems of RIP. It uses a link state routing protocol to generate a shortest-path tree, then builds up its routing table based on this tree. OSPF produces a more stable network because the participating routers act on network changes predictably and simultaneously, converging on the best route more quickly than RIP. Moreover, when several equal-cost routes to a destination exist, traffic can be distributed equally among them.

OSPF looks at more than just the simple hop count. When adding the shortest path to any node into the tree, the optimal path is chosen on the basis of delay, throughput and connectivity. OSPF utilizes IP multicast to reduce the amount of routing traffic required when sending or receiving routing path updates. The separate routing area scheme used by OSPF further reduces the amount of routing traffic, and thus inherently provides another level of routing protection. In addition, all routing protocol exchanges can be authenticated. Finally, the OSPF algorithms have been tailored for efficient operation in TCP/IP Internets.

OSPFv2 is a compatible upgrade to OSPF. It involves enhancements to protocol message authentication, and the addition of a point-to-multipoint interface which allows OSPF to run over non-broadcast networks, as well as support for overlapping area ranges.

Area Configuration – OSPF routers exchange information with other routers in their area to determine the shortest path to every destination. Each router in a common area should therefore have an identical map of their local network topology. At the top level, the largest area is known as an Autonomous System, and contains all the routers in your network. However, for large networks you should organize your OSPF routers into smaller contiguous areas to reduce the amount of routing information that has to be exchanged and to simplify network management.

When designing an OSPF network architecture, first create a backbone area to which all other areas are adjacent. Note that when you enable OSPF for any IP interface on the ES3627, it is assigned to the backbone by default (Area 0.0.0.0).

As a general rule, no area should not contain more than 50 routers. To create a new area, designate an Area ID that will be used by all of the other routers in this area, specify the area type as Normal, Stub, or NSSA (page 2-59 or 3-44), and then assign the ID to an interface (page 2-52 or 3-39). A Stub does not accept or send external routing information. Instead, it uses a single default route for destinations outside the area. Stubs further minimize the amount of routing data that has to be stored or exchanged with other areas. An NSSA (Not-So-Stubby Area) is similar to a Stub, except that it can import external route information into its area. Note that if there are not external routes into your network, then there are no advantages to configuring a Stub or NSSA.

Neighbors – Neighboring OSPF routers within a common area are found using Hello messages. These messages also list the other routers from which the originator has received hello messages. When a router finds its address in the hello messages received from another router, both routers initiate communications as neighbors.

Only after these routers successfully exchange and synchronize their routing tables, will they be considered fully adjacent (page 2-98 or 3-69). Routing information is only exchanged between adjacent neighbors.

Designated Router – A Designated Router (DR) and Backup Designated Router (BDR) are selected by

WGS3 Layer 3 Switch User’s Manual

- 236 -