Configuring Multi-Area OSPF with Area Types

1. The Case for Multi-Area OSPF

Imagine an enterprise network with hundreds of routers, all in a single OSPF area. Every time a link flaps or a router is added, an update (a Link-State Advertisement or LSA) is flooded to every single router. This leads to several problems.

To understand these challenges, let's start with a short video that explains why a single OSPF area doesn't scale.

By dividing a large OSPF domain into smaller areas, we can contain the impact of topology changes and optimize routing information. This hierarchical design is fundamental to OSPF's power and scalability.

2. Multi-Area OSPF Fundamentals

A multi-area OSPF design is built on a few key components and rules:

• Backbone Area (Area 0): This is the core of the OSPF network. All other areas must connect to the backbone area. Traffic moving between two non-backbone areas must pass through Area 0.
• Area Border Router (ABR): A router with interfaces in two or more areas, one of which is typically Area 0. ABRs are responsible for summarizing routes and advertising them between areas.
• Hierarchical Topology: The design resembles a hub-and-spoke model, with Area 0 as the hub and other areas as the spokes. This prevents routing loops and provides a logical structure.
• Router Roles: In a multi-area network, routers have specific roles. The video you just watched introduces these, but let's quickly review them:
  • Internal Router: All interfaces are in a single area.
  • Backbone Router: Has at least one interface in Area 0.
  • Area Border Router (ABR): Connects two or more areas.
  • Autonomous System Boundary Router (ASBR): Injects routes from other routing domains (e.g., BGP, EIGRP, or static routes) into OSPF.

The following segment of the same video explains these design principles and the role of the ABR in summarizing routes.

3. Understanding LSA Types: The Key to Area Design

To truly master OSPF area design, you must understand how OSPF shares information using Link-State Advertisements (LSAs). The primary function of different area types is to filter which LSAs are allowed to enter them. Before we dive into the area types themselves, let's review the most common LSAs.

Here is a summary of the LSAs you need to know for area design:

The core idea of special area types is to block the most "expensive" LSAs—Type 4 and 5—which advertise external routes that a simple leaf area may not need.

4. A Deep Dive into OSPF Area Types

With a solid understanding of LSAs, we can now explore the different area types. Each type represents a trade-off between route visibility and network efficiency.

The following video from Practical Networking provides an excellent, detailed walkthrough of each area type using a consistent topology. We will watch it in segments as we discuss each type.

a) Stub Area

A Stub Area is the first level of optimization. It's designed for areas that do not contain any ASBRs and do not need to know the specifics of external routes.

• Purpose: To reduce the LSDB size and routing table on routers within the area.
• Rules:
  • Blocks Type 4 and Type 5 LSAs from entering the area.
  • Routers in a stub area cannot see individual external routes.
• ABR Action: The ABR connected to the stub area automatically injects a Type 3 LSA for a default route (0.0.0.0/0) into the area. This allows routers in the stub area to reach external destinations without needing specific routes.

b) Totally Stubby Area (TSA)

A Totally Stubby Area is a Cisco-proprietary enhancement that provides even more optimization. It's ideal for very simple "leaf" or "spoke" areas where the only path out is via the ABR.

• Purpose: To maximally reduce the LSDB and routing table.
• Rules:
  • Blocks Type 3, Type 4, and Type 5 LSAs.
• ABR Action: The ABR injects a single Type 3 LSA for a default route. This is the only inter-area route that routers in a TSA will see. They will have intra-area routes plus one default route.

c) Not-So-Stubby Area (NSSA)

What if an area needs the resource savings of a stub area (blocking external routes from the core) but also needs to originate its own external routes? This is a common scenario in a branch that has its own local internet exit. This is where the Not-So-Stubby Area comes in.

• Purpose: To allow an area to behave like a stub area while still permitting local route redistribution.
• Rules:
  • Blocks Type 4 and Type 5 LSAs from entering from other areas.
  • Allows an ASBR within the NSSA.
  • This ASBR advertises its external routes using special Type 7 LSAs.
• ABR Action: The ABR translates the Type 7 LSAs from the NSSA into standard Type 5 LSAs and floods them into the backbone (Area 0) and other standard areas.

d) Totally NSSA

Finally, we can combine the concepts. The Totally Not-So-Stubby Area is a Cisco-proprietary version of NSSA that also blocks summary routes from other areas, just like a totally stubby area.

• Purpose: To provide maximum route optimization for a leaf area that also needs to originate external routes.
• Rules:
  • Blocks Type 3, Type 4, and Type 5 LSAs.
  • Allows an ASBR within the area, which uses Type 7 LSAs.
• ABR Action: The ABR injects a Type 3 default route and also translates Type 7 LSAs into Type 5 LSAs.

5. Summary and Real-World Use Cases

Now that we've covered the mechanics, let's consolidate this knowledge. The table below, from Network Journey, provides a perfect summary.

The article "OSPF Area Types Explained" provides a concise comparison and, most importantly, practical use cases for each area type. As a solutions engineer, matching the right technology to a business requirement is key.

Here's a summary of when to use each area type:

• Standard Area: Use for the backbone (Area 0) and any transit areas that need to pass full routing information between other areas.
• Stub Area: Use for a simple branch or remote site that connects only to the corporate network and doesn't need to know about external routes from BGP, etc.
• Totally Stubby Area: The best choice for a simple spoke in a hub-and-spoke topology. The branch router's routing table is minimized to local routes and a single default route.
• NSSA: Use for a branch or remote site that has its own connection to another network (e.g., a partner network or a secondary ISP) and needs to advertise routes from that network into the main OSPF domain.
• Totally NSSA: Combines the above two. Use for a remote site that needs to redistribute routes and wants the smallest possible routing table for routes from the corporate network.

Conclusion

In this lesson, we moved beyond single-area OSPF to understand the principles of scalable, hierarchical network design. We explored how dividing a network into multiple areas and using special area types can drastically reduce routing overhead and improve stability.

Key Takeaways:

• Multi-area OSPF solves scalability issues by creating smaller failure domains and enabling route summarization.
• All non-backbone areas must connect to the backbone, Area 0.
• The choice of area type is a design decision based on controlling LSA flooding.
• Stub areas block external routes (Type 4/5 LSAs).
• Totally Stubby areas block external and inter-area summary routes (Type 3/4/5 LSAs).
• NSSA areas block external routes but allow local redistribution using Type 7 LSAs.
• Totally NSSA areas block external and summary routes but still allow local redistribution.

Preview of the Next Lesson:

We've designed our areas to control LSA flooding. The next logical step in optimizing a multi-area OSPF network is to control the size of the routing table itself. In the next lesson, we will focus on how to configure and verify OSPF route summarization at ABRs and ASBRs, a powerful technique for creating smaller, more manageable routing tables.