OSPF or Open Shortest Path First is a router protocol implemented within larger autonomous system networks. It is fast replacing the Routing Information Protocol (RIP) which is an older routing protocol. This protocol is one of Interior Gateway Protocols (IGPs) that was designated by the Internet Engineering Task Force (IETF) . Unlike other Routing Protocols, here the entire routing table is not sent. Instead, the host using OSPF sends only the part that has changed. This protocol multicasts the updated information only in the case of any change occurrence. OSPF gives the user freedom to assign cost metrics to a particular host router so that certain paths can be given preference. This protocol also incorporates support for a variable network subnet mask, which makes it easy for a network to be subdivided. In this intermediate level tutorial, we introduce you to the different OSPF area types.
We assume that you know the basics of networking. If not, you may want to first learn about computer networking with this course.
Using the OSPF protocol, a host detects any change in the network. Without any delay, the host multicasts the information to the remaining hosts in the network. This is done so that all the hosts have the same routing table information. OSPF uses Dijkstra’s algorithm to compute the shortest path tree for each route. This algorithm is one of the shortest path first algorithms.
The OSPF network are subdivided into routing areas so that it can simplify the administration, optimize traffic and better utilize the resources. The protocol takes the responsibility to handle its own error detection and correction functions. Unlike the Routing Information Protocol (RIP) it does not use a TCP/IP transport protocol (if you’d like to understand the difference, check out this course on TCP/IP).
Advantages of OSPF vs RIP
OSPF permits better load balancing and does not restrict the hop count. Here routing authentication via different password authentication methods is allowed. This protocol allows the transfer and tagging of external routes to be put into an independent system. OSPF has better convergence than RIP as a result of instantaneous propagation of routing changes.
What are Areas in an OSPF Network?
The sub-domains of an OSPF network are called areas. An area consists of a number of OSPF networks, routers, and links. They share the same area identification. Also note that every router within an area should maintain a topological database of the area it belongs to. Since the router does not have the detailed information about its external network topology the size of its database is reduced.
Also understand that areas restrict the route information distribution. You cannot do route update filtering within an area. The link-state database (LSDB) of routers has to be synchronized and should be identical within the same area. Different areas can have route summarization and filtering between them. This protocol uses flooding process to exchange link-state updates between routers. Each change in routing info is communicated to all routers in the network. Areas are introduced to limit the explosion of link-state updates. To learn more about routing in OSPF, you can take this course.
Each OSPF network is divided into different areas, and must follow the rules below:
- Backbone area– It is essential to have a backbone area. The reason is that it merges a set of autonomous areas into one domain.
- Non-backbone area– Each of them are directly connected to the backbone area.
- Backbone areas cannot be partitioned- This must not take place regardless of the conditions like link or router failures.
An area ID represents each area. Usually area ids are displayed in IP address format. If there are many areas in your network, the backbone area will be named as “area 0.” The backbone that connects the areas in the network will be a contiguous area. If the backbone is partitioned there will be no access to the parts of the autonomous system. Further, the virtual links will have to be configured in order to prepare the partition. Learn more about setting up a switched IP network with this course.
Area border router
An area border router is defined as a router with interfaces in multiple areas. It exists in the OSPF boundary between two areas. Under any conditions, both sides of any link share the same OSPF area.
Autonomous system boundary router
An autonomous system boundary router (ASBR) displays external destinations across the autonomous system of this protocol. The redistributed routes from any other protocol into OSPF are known as external routes. In the databases of each router ,external link states are the most common link states.
OSPF Area Types
- Backbone Area
The backbone area is the integral part of an OSPF network. This area distributes the routing information between non backbone areas. The backbone connectivity is established and maintained by the configuring the virtual links.
- Stub Area
A stub area is defined as the area which does not display external routes. This results in further reduction of the size of the database. Alternately a default summary route (0.0.0.0) is put into in to the stub area. This is done so that one can reach these external routes. You do not have to define the stub areas if there are no external routes.
Stub areas get information about other networks that belong to other areas of the same OSPF domain. This occurs even though stub areas are protected from external routes. However, routers within the totally stubby area store their routing information within their database.
- Not-So-Stubby Areas (NSSA)
These areas are derived from the OSPF stub areas. They do not allow display of external links. Instead they leverage default routing to reach external destinations. Consequently, NSSAs are located at the periphery of an OSPF routing domain. It is possible for NSSA to import external routes into the OSPF routing domain. This way transit services is provided to small routing domains which do not belong to the OSPF routing domain.
A host of vendors (Cisco, Allied Telesis, Juniper, Alcatel-Lucent), also support the two extensions below to stub and NSSA area. Even though not part of RFC, they are regarded by many to be standard features in OSPF implementations.
- Totally Stubby Area (TSA)
This is similar to a stub area. Here inter-area(IA) routes are not translated into totally stubby areas. To get traffic routed outside of the area the default route method is used. This is the sole Type-3 LSA displayed in this area. As there is only a single route out of the area, less routing decisions are required. This in turn reduces the utilization of the system resources.
- NSSA Totally Stubby Area
The totally stubby NSSA inherits the attributes of a TSA (Totally stubby area). This is along with an NSSA’s standard features. Here type 3 and 4 summary routes are not send into this OSPF area type. Here you can also declare an area as both totally stubby and not so stubby. As a result the area will only get the default route from area 0.0.0.0. It also has an autonomous system boundary router (ABSR). This router takes external routing information and puts it into the local area. This operation is also done from the local area into area 0.0.0.0.
- Transit Area
This area has two or more OSPF border routers. Transit areas passes network from one adjacent area into another. However this area is neither the originator nor the destination of such traffic.
We’ve tried to covered the basics of OSPF area in this article. However, there’s much more to it. You may want to check out this course OSPF for Enterprise and Service Provider networks to get a better understanding of it.