What Multi Area Ospf Computer Science Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

We are assigned with the task of implementing a network consists of multi area OSPF including NSSA with external network and EIGRP redistribution. In this paper, we will discuss what multi area OSPF is and EIGRP, especially the portion related to our project. We will give a brief overview how multi area OSPF works and how NSSA behaves with external network like EIGRP in our case. We will implement these all within an Autonomous system and provide detailed configurations.


A routing protocol which requires routers to keep track of their reachable destinations or floods new updates into network, becomes infeasible as the size of the network increases. Though there are many approaches to minimize these overheads but routing hierarchy is used to provide scalability and manageability to a large network [1].

The network hierarchy restricts the network topology changes and routing updates to a specific area, thus, reducing the size of routing table and other topology tables [2]. In addition to this, routing hierarchy also makes it easier to diagnose any fault occurred in the network.

OSPF uses several types of link state advertisements (LSAs), each of them has it's own descriptions in different areas. In our topic we will discuss about type-7 LSA used in Not-so-stubby-area. Type-5 LSA (external route) is not accepted in NSSA area, so the ASBR of NSSA generates this type-7 LSA to be allowed in NSSA area and translated back to type-5 LSA by the NSSA Area Boundary Router to be propagated to type-5 LSA capable areas[6].


Open shortest path first (OSPF) is an open standard link state routing protocol and widely used in modern networks. Default administrative distance of OSPF is 110. Router running ospf has a full map of the network and the bandwidth information of each individual links to the destination network. Ospf uses routing decisions based on the loop free shortest path available. it is classless routing protocol supporting VLSM. Unlike distance vector protocols, ospf doesn't send periodic updates. It responds quickly to a change occurs in the network and updates the neighbors about that change. It also sends periodic updates at each 30 minutes known as link state refresh.

Ospf maintains three type of tables:

Neighbor table: it contains information about recognized neighbors, it is also known as adjacency table.

Topology table: contains information of routers and the status of their attached links, also known as Link State Data Base LSDB.

Routing table: Know as Forwarding Database, contains the list of best paths to the destination.

OSPF has many advantages over Distance Vector Protocols. It converges very fast, it uses lower bandwidth by sending updates only when a change occurs in the topology and it is also very scalable supporting very large networks.


If OSPF is implemented in a large networks using a single area, the size of the Routing table, Neighbor table and LSBD will dramatically increase. A change occurs in a link will be updated to all the routers in that area. A topology change will cause all the routers to re-establish the spf tree and form adjacencies in that particular area. If an area consists of large network, these calculations will consume much memory and cpu process. The larger the databases of ospf, the more is the memory and time required for processing, neighbor discovery and form adjacencies.

To solve these limitations, OSPF autonomous system is divided into different smaller internetworks called Areas. "Area 0" is considered as Backbone Area. All other areas must be connected to Area 0, to enable inter area routing. With this approach routing between different areas will still occur but will decrease the processing burden over the cpu by limiting the algorithm for spf re-running and re-calculating the database to that individual area where changes have been occurred. Which increase the scalability and trouble shooting of ospf.


OSPF has four router types:

Backbone Router: A Router with minimum one interface connected to area 0 is called Backbone Router.

Internal Router: A router which has all its interfaces configured within the same area.

Area Border Router: ABR is a router which has interfaces configured in different areas.

Autonomous System Boundary Router: ASBR is router that has at least one link connected to another Autonomous System.

OSPF has many advantages over Distance Vector Protocols. It converges very fast, uses lower bandwidth by sending updates only when a change occurs in the topology or in the link and it is also very scalable supporting very large networks.


In order to establish neighbor relationship and form adjacency, ospf exchanges five packet types with other routers running ospf in the network. But before exchanging the link state information with other ospf, Designate Router (DR) and Backup Designate Router (BDR) must be elected. Once DR and BDR have been elected, router that added after this election will only establishes its adjacency with DR and BDR

The five packets type which exchange during neighbor discovery are:

Hello Packet: in order to detect and maintain neighbors, ospf sends Hello Packet to other routers running ospf.

Database Description Packet: it has the information about the link-state database of ospf.

Link State Request: it requests some particular information about the link state database of the neighbor router.

Link State Update: It sends link state updates (Link state Advertisements) being asked to the neighbor router.

Link State Acknowledgement: the neighbor router sends back an acknowledgment of successful receipt of the requested information.

OSPF packet 4, i.e. Link State Advertisements has further more LSA types

these are described here [3].


(Router Link Entry, O-OSPF)

It is generated by each router in the same area which describes the cost and the status of the link. This information is sent to other routers in a specific area.


(network link entry, O-OSPF):

It describes the set of routers that exist in a particular area. This information is generated by the DR and is only for the area which exists the network.


(summary link entry, IA-OSPF inter area):

It is originated by the Area Border Routers ABRs, which describes routes to the network of local area and is sent to the Backbone area.


(summary link entry, IA-OSPF inter area):

It is also generated by the ABRs, LSA type 4 shows information about reachability to the Autonomous system boundary router ASBR. This link information is not sent to the the totally stubby areas.


(E1-OSPF external type1):

LSA 5 messages are generated by ASBRs, which describe external routes information of destination network that is not in the same autonomous system. Shown in the routing table as E1 and E2.

These routes information is not forwarded to stub and totally stubby areas.


(Multicast OSPF)

It is used for multicast purpose by enabling the ospf routers to use their link state databases and build multicast distribution tree to forward multicast data.


NSSA External Link Entry:

LSA type 7 messages are generated by the ASBRs which connect to Not-So-Stubby Area NSSA. These messages are sent through out NSSA where they originated. ABR translates LSA 7 into LSA type 5 to flood into other areas. N1 and N2 denotes NSSA routes.

LSA 8:

This is a special type of LSA, used for internetwork between OSPF and BGP [4]

LSAs 9, 10, 11:

These LSAs are designed for the future development of ospf specific applications [3].


In order to handle different kind of traffics, Ospf uses several area types for different purposes. These areas are categorized as,

Standard or Normal area: this is the area which accepts all information regarding link updates, summarization and external routes.

Backbone Area: Also known as Transit Area, is the central area, where all other areas must be connected to exchange route information. It is denoted as 'Area 0" and has similar features that exist in Standard Area.

Stub Area (SA):

Stub area is area which has only one exit point from that area. It is not allowing external routes information to enter in it's AS. Meaning type 5 and type 4 LSAs are not accepted. For route to external network, ABR distributes a default route into the Stub area, which is then used for communication to external networks.

Totally Stubby Area (TSA):

An area is considered to be Totally Stubby Area, when there is only one exit path to the external network [3] from that particular area (single ABR) or it does not have to chose any optimal path for external networks. Totally stubby area doesn't accept external route and summary route information to enter in the autonomous system where it exists. LSA5, LSA4 all are blocked by the Area Boundary Router ABR. It also blocks type 3 LSAs but allowing only a single LSA to distribute a default route by ABR in Totally Stubby Area [2]. This default route is used by TSA to reach networks external to the autonomous system. Stub Area and Totally Stubby Area, save the memory size and processing time of the cpu [3]. Totally stubby area can't become ASBR but only in that case if ABR is also configured as ASBR.

Fig.1, OSPF areas [3]

Not-So-Stubby-Area NSSA:

NSSA is relatively a new enhancement to ospf, which has the benefits same to stub area and totally stubby area but it also accepts external routes to be advertised inside the autonomous system. NSSA allows ASBR to be configured, which was not possible in stub area and totally stubby area. ASBR generates type-7 LSAs to distribute external route into NSSA. This route is propagated through out NSSA but blocked by the ABR.

To accept limited exchange of external route information across NSSA, Area border router translates type-7 LSAs received from NSSA into type-5 LSAs. The ABR will then send these type-5 LSAs into other areas which accept type-5 LSAs. The NSSA area border router can be configured with an address range to represent many type-7 LSAs with a single type-5 LSA [5].

NSSA area border router may generate default type-7 LSAs with an ip address of inside not-so-stubby-area. A default route is required because NSSA does not receive all routing information and it should have a default route to reach external destinations [5].


Enhanced interior gateway protocol (EIGRP):-

EIGRP is Cisco proprietary hybrid protocol. It is classless routing protocol which uses complex metric. The key features of EIGRP are, convergence is fast, Supports VLSM, and Supports IP, Apple talk and IPX. EIGRP has Layer 2 Support, uses Multicasting instead of broadcasting and also has Authentication Support.

EIGRP mainly works on four technologies.

Neighbors maintenance and discovery.

EIGRP sends hello packets for discovery.

The RTP (reliable transport protocol).

RTP controls sending, Acknowledging and tracking EIGRP messages.

Diffusing Update Algorithm (DUAL).

Dual algorithm is used to determine best loop free route.

Protocol independent modules (PDM).

These are plug-ins for IPX, IP and Apple Talk.

For management purposes the EIGRP uses three tables. The neighbor table, topology table and routing table.

There are five main packets types used in EIGRP.

(a) Hello (b) Update (c) Query (d) Reply (e) Ack

Hello packet identifies neighbors, Update packet sends route information, Query packet requests for route information, Reply packet responds to query and Ack packet is used for Acknowledgement.

EIGRP neighbor discovery and route exchange:-

On starting EIGRP sends hello packets to build a neighbors table. Neighbors are the attached routers having same AS number and k value.

Communication between two EIGRP routes is as under

Step1. Router A sends hello.

Step2. Router B sends back hello and Update.

Step3. Router A acknowledges update.

Step4. Router A sends update.

Step5. Router B acknowledges.

For keepalives EIGRP neighbor uses hello packets between them. The neighbor will be considered lost if no hellos received with in three hello periods. The neighbors table can be seen by the command "show ip eigrp neighbors".

EIGRP route selection is done by the comparison of the Advertise distance (AD), feasible distance (FD), K value and Bandwidth to a route.

Bandwidth is main consideration in metric. DUAL algorithm chooses best path by looking AD and FD.

EIGRP redistribution:-

"When a default route, or in fact any

route, is made known to a routing protocol that did not natively discover the route, it is known as redistribution".[a]

Redistribution imports routs known to other protocols, then routing known to one protocol is distributed in other routing protocol back and forth. These routes are then advertised by current routing protocol.

In EIGRP the internally learned routes are more trusted then the externally learned routes. The administrative distance (AD) shows us the priority of the routes such as, EIGRP (summary routes) has AD of 1, EIGRP (internal) has AD of 90 and EIGRP (external) has AD of 170.




[4] http://Cisco.netacad.net

Reference:[5] http://www.faqs.org/rfcs/rfc1587.html