Preferably using a whiteboard, start with the hierarchy of OSPF — a two-level model — and draw a diagram like the one in Figure 2-9.
Discuss having a backbone Area 0 (or 0.0.0.0) and that all areas must connect to the backbone area. Other area types include the following: stub area (an area that does not receive AS external routes); total stubby area (an area that does not allow summary routes or external routes); and not-so-stubby area (an area that can import AS external routes and send them to the backbone area, but will not receive AS external routes from the backbone or other areas). You want to include the type of routers contained in the hierarchy: internal routers, area border routers (ABR), backbone routers, and autonomous system boundary routers (ASBR). Mention that the shortest path first (SPF) calculation is performed independently on each area. You should state that the time to converge is faster than distance-vector routing protocols (DVRPs) such as RIP. Include a brief statement on the low bandwidth requirement for LSAs. Also mention support for classless routing, Variable-Length Subnet Masking (VLSM), authentication, and multipath statements.
Describe the OSPF algorithm and generally how it works: Changes in the network generate LSAs, routers exchange the LSAs, and each router builds and maintains its own database. So if the network is in a steady state, there will be refresh LSAs only every 30 minutes. You want to cover the five types of routing protocol packets: Hello, Database description, Link-state request, Link-state update, and Link-state
acknowledgment. Hello packets are multicast on 224.0.0.5 and routers use them to form adjacency relationships.
You want to cover the types of Link State Advertisements (LSAs): Router link (LSA type 1), Network link (LSA type 2), Network summary (LSA type 3), ASBR (LSA type 4), External (LSA type 5), and NSSA external (LSA type 7). Do not neglect a discussion on IPv6 and that OSPFv3 supports it. You might want to discuss briefly how vendors implement OSPFv3 using a ships-in-the-night approach to support both v3 and v2 simultaneously. OSPFv3 distributes IPv6 prefixes and uses the same interfaces
and nearly the same LSA types as OSPFv2. It uses the same methods for neighbor discovery and adjacency forming. The only differences are that OSPFv3 has to use a network link rather than a subnet.
And there can be multiple instances of OSPFv3 on a given link. You should mention that the topology in OSPFv3 is a bit different as well — using a router ID and Link ID. And because OSPFv3 uses links, there’s a new Link LSA type as well as an Intra-Area Prefix LSA for the IPv6 prefixes.
To fully answer the question, you have to go through the process of neighbor finding and adjacency creation. Routers sharing a common network segment or link will become neighbors using the Hello protocol. Routers send Hello packets out each interface with the multicast address of 224.0.0.5. When a router sees its own primary address in a Hello packet from another router, the routers are then neighbors. As
neighbors, the routers have to agree on the following things: the area-id (of the area they belong to); a preshared password (for authentication); hello and dead intervals (how often hello packets are sent and how long to wait to for a neighbor’s hello); and a stub area flag (whether the router is in a stub area). Neighbor routers then form adjacencies. When the routers exchange their databases, they are adjacent. To limit the volume of information exchanged on a network segment, routers go through an election process. This election nominates a designated router (DR) and a backup designated router (BDR). The DR is the sole source for updates on the segment. All other routers on the segment exchange route information with the DR and BDR. Area Border Routers (ABRs) collect all the routes for the area and combine/
summarize them into a single advertisement to the backbone area (inter-area route summarization). The backbone routers then forward these summarized routes. External route summarization may occur as well when distributing the routes to another protocol.
For more information on interior routing protocols, check out Routing TCP/IP, Volume 1 (2nd Edition) (CCIE Professional Development) by Jeff Doyle and Jennifer Carroll (Cisco Press. 2006).