In order to configure and troubleshoot OSPF over a network, you must have a good understanding of the underlying network topology. The neighbor discovery mechanism, election of Designated Router (DR) and Backup Designated Router (BDR), and update flooding depend on it. The underlying Layer 2 topology can be one of these:
Broadcast multi-access (for example, Ethernet) - A Broadcast network is one which allows broadcast or multicast packets to be sent over the network and hence every device able to communicate directly with every other device in the segment. A multi-access network is a network that connects more than two devices.
Point-to-Point (such as serial links with Point-to-Point and High-Level Data Link Control (PPP/HDLC)) - point-to-point networks also allow broadcast or multicast packets to be sent over the network, and these networks connect only two devices on the segment.
Non-Broadcast multi-access (NBMA), such as Frame Relay - These networks do not support broadcasts or multicasts, but can connect more than devices and are multi-access in nature.
Point-to-Multipoint - This is a collection of point-to-point links between various devices on a segment. These networks also allow broadcast or multicast packets to be sent over the network. These networks can represent the multi-access segment as multiple point-to-point links that connect all the devices on the segment.
When OSPF is run on a network, two important events happen before routing information is exchanged:
Neighbors are discovered using multicast hello packets.
DR and BDR are elected for every multi-access network to optimize the adjacency building process. All the routers in that segment should be able to communicate directly with the DR and BDR for proper adjacency (in the case of a point-to-point network, DR and BDR are not necessary since there are only two routers in the segment, and hence the election does not take place).
For a successful neighbor discovery on a segment, the network must allow broadcasts or multicast packets to be sent.
In the broadcast multi-access Layer 2 topology, broadcasts are supported; therefore, a router that runs OSPF can discover OSPF neighbors automatically and elect any router as DR and BDR since any device can talk to all other routers in that broadcast segment.
In a point-to-point topology, neighbors are discovered automatically since neighbors are directly connected to each other through a point-to-point link, and broadcast or multicast packets are forwarded over the network; however, the DR and BDR election does not take place as explained earlier.
In an NBMA network topology, which is inherently nonbroadcast, neighbors are not discovered automatically. OSPF tries to elect a DR and a BDR due to the multi-access nature of the network, but the election fails since neighbors are not discovered. Neighbors must be configured manually to overcome these problems. Also, additional configuration is necessary in a hub and spoke topology to make sure that the hub routers, which have connectivity with every other spoke router, are elected as the DR and BDR. Alternatively, you can change the configuration on the NBMA interface to make OSPF believe that it is another network type that does not have these problems.
The correct configuration is necessary for the proper operation of OSPF.
Frame Relay subinterfaces can run in two modes:
Point-to-Point - When a Frame Relay point-to-point subinterface is configured, the subinterface emulates a point-to-point network and OSPF treats it as a point-to-point network type.
Multipoint - When a Frame Relay multipoint subinterface is configured, OSPF treats this subinterface as an NBMA network type.
Cisco IOS software uses the ip ospf network command to allow the flexibility run OSPF on an interface in different modes:
"ip ospf network {broadcast | non-broadcast | {point-to-multipoint [non-broadcast] | point-to-point}}"