Repairing a Discontiguous OSPF Backbone

You already know the first job of a virtual link: attaching an area that never touched Area 0. There is a second case, and this one hits the backbone itself. Here, Area 0 is not missing, it is broken in two.

The Healthy Backbone

Start from a working design. Area 0 is present on both sides of the topology, connected through R10 at the top:

Figure 1 - The backbone is contiguous

As you can see, the topology forms a ring: R1, R2, R10, R5, and R4 are all connected inside Area 0.

The backbone is contiguous. Contiguous simply means in one piece: starting from any Area 0 router, you can reach every other Area 0 router using only Area 0 links. As long as this is true, everything works.

The Backbone Partition Event

Now the link between R10 and R5 goes down:

Figure 2 - The backbone link fails

One cable, and Area 0 is cut in two pieces: the backbone is discontiguous.

Discontiguous means the opposite: Area 0 is no longer in one piece. The left half and the right half are separated by Area 1, and no Area 0 link connects them anymore.

Figure 3 - Area 0 is split in two

Each half still floods its LSAs and converges on its own. But the two halves never exchange their routes again: the routers on the left can no longer reach the networks on the right.

Why this confuses engineers: Area 1 is still up and forwarding traffic. The physical path exists, but the OSPF path does not because OSPF strictly forbids inter-area routing through a non-backbone area to reach another Area 0 segment.

The fix uses the same tool as before, but for a different job.
Instead of attaching an orphan area, the virtual link stitches the broken backbone back to itself across Transit Area 1.