When studying link-state routing protocols, Open Shortest Path First (OSPF) stands out as one of the most important and widely deployed technologies. OSPF is commonly used in enterprise networks and service provider environments due to its fast convergence, scalability, and open-standard design. A solid understanding of OSPF is therefore essential for anyone involved in modern IP network design or operations.

But what exactly is OSPF, and why is it so effective in large networks?

OSPF Areas

The full name of OSPF is Open Shortest Path First. The term Open indicates that OSPF is a public, open standard, allowing any vendor to implement it freely. Shortest Path First (SPF) refers to the algorithm used by OSPF to calculate optimal routing paths.

As an open-standard, link-state routing protocol based on the SPF algorithm, OSPF has the following characteristics:

• It is an open protocol.
• It converges quickly after topology changes.
• It supports incremental and triggered updates.
• It supports subnet masks and VLSM.
• It supports both IPv4 (OSPFv2) and IPv6 (OSPFv3).
• It maintains three main tables:
  • Neighbor table
  • Topology table (link-state database)
  • Routing table

At first glance, OSPF may appear similar to EIGRP, particularly in terms of fast convergence and efficient updates. This naturally raises an important question: aside from using different routing algorithms, what distinguishes OSPF from EIGRP, and what advantages does OSPF offer?

The most significant feature that sets OSPF apart is its support for areas. This concept is not only OSPF’s defining characteristic, but also the primary reason it scales effectively in large networks.

Why Areas Matter

To understand the importance of areas, consider a historical analogy. During the feudal period in Europe, kings governed vast territories by delegating authority to regional lords. For example, the King of France divided his land and entrusted portions to nobles.

Why would a ruler willingly give up direct control? The answer lies in scalability. As territories expanded, centralized administration became inefficient. Communication was slow, administrative complexity increased, and direct control became unsustainable. Dividing the territory into hierarchical regions simplified governance and improved stability.

Computer networks face a similar challenge. As a network grows, the amount of routing information increases rapidly. If a routing protocol cannot divide a large network into smaller, manageable sections, it will struggle to scale.

OSPF solves this problem by allowing the network to be divided into areas. This design reduces routing overhead, limits the impact of topology changes, and enables OSPF to scale far beyond what distance-vector protocols such as RIP—or even flat routing protocols like EIGRP—can comfortably support.

OSPF Area Structure

Figure 1 shows an example OSPF topology. The network is divided into three areas: Area 0, Area 1, and Area 2. Both Area 1 and Area 2 connect directly to Area 0.

In OSPF, a router with interfaces in more than one area is known as an Area Border Router (ABR), as shown in Figure 2.

ABRs play a critical role in OSPF. They maintain separate link-state databases for each connected area and can summarize routes at area boundaries. By advertising summarized information rather than detailed topology data, ABRs significantly reduce routing overhead and prevent topology changes in one area from propagating throughout the entire network.

Compliance: All OSPF areas must connect to Area 0, known as the backbone area.
If a physical connection is not possible, a virtual link must be used to maintain logical connectivity to Area 0 and prevent routing loops.

How OSPF Works

With the concept of areas established, we can now examine how OSPF operates.

At a high level, OSPF learns routing information through four main steps:

1. Routers discover neighbors and establish bidirectional adjacencies.
2. Adjacent routers exchange link-state information.
3. Each router builds an identical link-state database for its area.
4. Each router independently runs the SPF algorithm and installs the best routes in its routing table.

Steps 2 through 4 are illustrated in Figure 3.

Note: In OSPF, link-state information is carried in packets that contain one or more Link-State Advertisements (LSAs). These LSAs are flooded unchanged throughout an area.

Adjacency and Hello Packets

Neighbor discovery is the foundation of any link-state routing protocol. When two OSPF routers detect each other, confirm bidirectional communication, and satisfy specific requirements, they form an adjacency. Once adjacent, routers synchronize their link-state databases.

Adjacencies must also be maintained, not just established. Routers continuously verify neighbor reachability, ensuring that outdated routing information does not persist.

OSPF performs neighbor discovery and maintenance using Hello packets.

Conceptually, a Hello packet is similar to saying “hello” during a phone call. It confirms presence, identity, and continued communication. In OSPF, Hello packets serve two purposes:

• Discovering neighbors
• Verifying that existing neighbors are still reachable

OSPF routers periodically send Hello packets to multicast address 224.0.0.5. Each Hello packet includes the router’s router ID, which uniquely identifies the router within the OSPF domain.

Router ID Selection

OSPF uses router IDs because hostnames are not guaranteed to be unique. A router ID is represented as a 32-bit value in IP address format and must be unique.

By default, Cisco routers select:

1. The highest IP address on a loopback interface
2. If no loopback exists, the highest IP address on an active physical interface

Administrators can also manually configure the router ID if needed.

Hello Packet Parameters

Hello packets also contain parameters that must match for adjacency formation, including:

• Hello and Dead intervals
• Authentication settings
• Network mask
• Area ID

If any of these parameters do not match, the Hello packet is discarded.

However, matching parameters are a necessary but not sufficient condition for adjacency formation.

OSPF Network Types

Whether routers form adjacencies with each other also depends on the network type.

Point-to-Point

A point-to-point network has exactly two endpoints. In this environment, if Hello parameters match, the routers always form an adjacency. Failing to do so would fragment the OSPF domain.

Broadcast

Broadcast networks, such as Ethernet, may contain many routers. Forming adjacencies between every router pair would be inefficient. To address this, OSPF uses a Designated Router (DR) and Backup Designated Router (BDR).

Only the DR and BDR form full adjacencies with all routers. Routers that are neither DR nor BDR—called DROTHER routers—do not form adjacencies with each other.

Figure 4 shows an Ethernet broadcast network, and Figure 5 illustrates the resulting adjacency relationships.

All routers send Hello packets to 224.0.0.5. DROTHER routers send link-state updates to 224.0.0.6, which is listened to only by the DR and BDR. The DR then floods updates to all routers.

Point-to-Multipoint

Point-to-multipoint networks are treated as collections of point-to-point links. No DR or BDR election occurs, and routers form adjacencies directly with each other.

Summary: OSPF network type determines adjacency behavior and whether DR/BDR elections occur, directly impacting scalability and flooding efficiency.

Election of DR and BDR

The DR and BDR election is based on:

1. OSPF priority (highest wins)
2. Router ID (used as a tie-breaker)

The default priority is 1. A priority of 0 prevents a router interface from participating in the election. The router with the highest priority becomes the DR; the second-highest becomes the BDR.

Important: DR and BDR roles apply to interfaces, not to the router as a whole.

Shortest Path First (SPF) Algorithm

To understand SPF, imagine a traveler in Paris using a rail timetable to find the shortest routes to other European cities. From Paris, there may be multiple paths to each destination. By comparing options, the traveler selects the shortest route and records it, eventually forming a tree of optimal paths.

Each OSPF router does the same. Using its link-state database, it independently runs Dijkstra’s SPF algorithm to build a shortest-path tree rooted at itself.

Routers that belong to multiple areas—ABRs—maintain a separate SPF tree for each area.

OSPF Cost Metric

OSPF does not mandate a specific metric. Cisco implementations use bandwidth:

For example:

• A 10 Mbps link has a cost of: $$ \frac{10^{8}}{10,000,000} = 10 $$
• A 100 Mbps link has a cost of: $$ \frac{10^{8}}{100,000,000} = 1 $$

Lower cost paths are preferred. The lowest-cost routes are installed in the routing table.

OSPF Configuration Commands on Cisco Routers

OSPF configuration is slightly more complex than EIGRP due to the use of areas, but the overall logic remains straightforward.

Enabling OSPF

router ospf process-id

The process ID identifies the local OSPF process. It has local significance only and does not affect neighbor relationships.

Advertising Networks

network 10.0.0.0 0.0.0.255 area 0

This command advertises the 10.0.0.0/24 network into Area 0. Although areas are conceptually dotted-decimal, Cisco commonly uses decimal notation.

Router ID Configuration

router-id 1.1.1.1

Interface-Level Configuration

• Set cost: ip ospf cost 10
• Set DR/BDR priority: ip ospf priority 100

OSPF Configuration Laboratory

Standard OSPF Configuration

The laboratory topology is shown in Figure 6.

All networks are in Area 0. Router IDs are configured as follows:

• R1: 1.1.1.1
• R2: 2.2.2.2
• R3: 3.3.3.3
• R4: 4.4.4.4
• R5: 5.5.5.5

Configuration commands are shown in Figures 7 through 11.

OSPF Configuration Inspection

The show ip ospf neighbor command verifies adjacency relationships. On R2, adjacencies exist with R1, R3, and R4.

• FULL/- indicates a point-to-point link
• FULL/BDR indicates a broadcast link with DR/BDR election

A FULL state confirms that LSAs are being exchanged.

Viewing OSPF Tables

• Topology table (LSDB): show ip ospf database

  

• Routing table: show ip route ospf

  

• Interface details: show ip ospf interface

  

• Routing protocols overview: show ip protocols