Open Shortest Path First
ospfd is an Open Shortest Path First
(OSPF) daemon which manages routing tables. This implementation supports
OSPF version 2, thus it is only capable of maintaining IPv4 routing
OSPF is an interior gateway protocol designed to supersede RIP. It has several advantages over RIP. For instance, every router has an understanding of the complete network topology. Response to changes in the network is faster. Furthermore, failure detection is improved.
The OSPF daemon maintains a Link State Database (LSDB) containing information about routers and networks within an Autonomous System (AS).
Dijkstra's shortest path first algorithm is used to compute a Routing Information Base (RIB) using the LSDB as input. The Forwarding Information Base (FIB), a.k.a. the kernel routing table, is updated with information from the RIB.
OSPF routers discover one another automatically via OSPF hello packets. OSPF routers communicate via two multicast groups: 188.8.131.52 (all Shortest Path First routers) and 184.108.40.206 (all Designated Routers). OSPF runs directly on top of IP and uses neither TCP nor UDP. IP protocol number 89 is reserved for OSPF.
All routers in an OSPF network spend most of their time keeping each others' LSDBs in sync. All routers must have the same information in the LSDB at all times. Every time the LSDB is updated the RIB is updated; if needed the FIB is also updated.
In a multi-access network such as Ethernet, it is unfeasible for all routers to synchronize their LSDB with all other routers in the network. In such networks a Designated Router (DR) and a Backup Designated Router (BDR) are elected. The DR's responsibility is to synchronize with all routers; the BDR will not do much until the DR fails. The first router in a network is automatically elected DR, the second router BDR. All routers have a FULL adjacency with the DR and the BDR. Routers with FULL adjacency exchange information about their LSDBs. A router not elected either DR or BDR will have 2-WAY adjacency with all routers but the DR and BDR. Routers with 2-WAY adjacency recognize that they know each other, but do not exchange information about their LSDBs. If a DR or BDR fails another router is elected DR or BDR and all routers form FULL adjacencies with the newly elected DR or BDR.
When routers are connected via point-to-point links, DR and BDR election is skipped since only two routers are connected to the link.
To limit the impact changes in the network have on the LSDB it is possible to segment an OSPF network into areas. Area 0.0.0.0 (a.k.a. the backbone area) must always be present. Routers can be configured as Area Border Router (ABR), being part of multiple areas. Every area must have direct access to the backbone area. ABRs not directly connected to the backbone area need to establish a virtual link to a router in the backbone area.
AS Border Routers (ASBR) are connected to an OSPF network and other external networks via BGP, RIP, or static routing, and provide connectivity to networks outside the AS.
ospfd is usually started at boot time, and
can be enabled by setting the following in
See rc(8) and rc.conf(8) for more information on the boot process and enabling daemons.
ospfd can be controlled with the
The options are as follows:
- Define macro to be set to value on the command line. Overrides the definition of macro in the configuration file.
- Do not daemonize. If this option is specified,
ospfdwill run in the foreground and log to stderr.
- Specify an alternative configuration file.
- Configtest mode. Only check the configuration file for validity.
- Use an alternate location for the default control socket.
- Produce more verbose output.
- UNIX-domain socket used for communication with ospfctl(8).
J. Moy, OSPF Version 2, RFC 2328, April 1998.
A. Retana, L. Nguyen, R. White, A. Zinin, and D. McPherson, OSPF Stub Router Advertisement, RFC 3137, June 2001.
ospfd program first appeared in
Virtual links are currently not available in