Understanding OSPF Route Selection for CCIE Enterprise Certification
When preparing for the CCIE Enterprise certification, understanding how OSPF selects the best route is one of those topics that can make the difference between simply knowing commands and actually understanding how the protocol behaves. OSPF is widely used in enterprise networks because it adapts quickly to topology changes, supports hierarchical designs, and makes intelligent routing decisions.
But OSPF route selection is more than just choosing the route with the lowest metric. Several factors come into play, including route type, administrative distance, cost, area structure, and the route's destination. Knowing the exact decision process helps you troubleshoot routing problems confidently and handle the scenario-based questions often associated with advanced Cisco certification preparation.
Why route selection trips people up
OSPF does not pick a path by comparing cost alone. It runs a sequence of checks, and cost only matters once the earlier checks come out equal. Candidates that skip straight to cost end up confused when a "more expensive" route wins. After years of troubleshooting enterprise networks and coaching engineers through lab prep, I've seen this misunderstanding cause more wrong answers than any other single topic in OSPF.
The order of decisions
Before OSPF gets involved at all, the router applies the rules every routing table follows. The longest prefix match wins first, so a /26 beats a /24 no matter what protocol installed it. If the same prefix comes from several protocols, administrative distance decides, and OSPF defaults to 110.
Only then does OSPF's own logic apply. It ranks routes by type before it ever looks at cost. The preference runs like this:
Intra-area routes (O)
Inter-area routes (O IA)
External type 1 routes (E1 or N1)
External type 2 routes (E2 or N2)
That ranking is the heart of this topic. An intra-area route with a cost of 500 will beat an inter-area route with a cost of 20. OSPF trusts what it learned directly inside the area more than what an ABR summarized for it.
A quick reference table
Keep this table in your head, because the lab loves to hand you a redistribution scenario and ask which path a router will choose.
How cost is calculated
Cost on a Cisco interface is the reference bandwidth divided by the interface bandwidth. The default reference is 100 Mbps, which means FastEthernet, Gigabit, and 10 Gigabit links all end up with a cost of 1. That is rarely what you want in a modern network. You can fix it with auto-cost reference-bandwidth under the OSPF process, but set it the same on every router, or you will create inconsistent path choices that are painful to track down.
You can also override cost directly with ip ospf cost on an interface, which is often the cleaner tool when you need to steer traffic in a specific way.
E1 versus E2 in practice
External routes cause the most head-scratching, so spend time here. An E2 route carries only the metric set at redistribution, and the cost to reach the ASBR is ignored unless two E2 routes tie. That is why E2 is the default and why it can lead to surprising results. If you have two ASBRs advertising the same external prefix with the same seed metric, the router will quietly pick the one closer to it. If the seed metrics differ, the closer ASBR can lose.
An E1 route adds the internal cost to the ASBR on top of the external metric, so the total reflects the whole journey. In networks with several exit points, E1 usually gives you more intuitive behavior. The trade-off is that you need to be deliberate about setting it during redistribution.
NSSA areas add a small wrinkle. Type 7 LSAs are translated to Type 5 at the ABR, and when several NSSA ABRs exist, the one with the highest router ID does the translating. Knowing that detail has saved me more than once when an external route showed up from an unexpected direction.
Equal cost paths and load sharing
When two or more paths tie on type and cost, OSPF installs them all, up to the configured maximum. Many platforms default to four, and you can raise it with maximum-paths. Remember that this is per-destination or per-flow load sharing depending on the forwarding mechanism, not a perfect split of bandwidth.
What to check when the result looks wrong
When a route choice surprises you, work through it methodically:
Confirm the prefix length and look for a more specific route hiding somewhere
Check whether another protocol is winning on administrative distance
Read the route code in show ip route before you look at the metric
Use show ip ospf database to see which LSA type is actually advertising the prefix
Run show ip ospf border-routers to verify the path toward an ASBR or ABR
Nine times out of ten, the answer is visible in the LSA type or the route code, not the metric.
Tips for CCIE Enterprise preparation
Build the topology yourself rather than reading about it. Create a multi-area design, redistribute from another protocol at two points, and then change costs and metric types one at a time while watching the routing table. Predict the result before you press enter. That habit of predicting first is exactly what the lab rewards, because you are working against the clock and cannot afford to guess.
Also practice the loop prevention rules around area 0. An ABR only accepts inter-area information from the backbone, which explains why traffic sometimes takes a path through area 0 that looks longer on paper. Once that clicks, a lot of odd behavior starts making sense.
Final thoughts
OSPF route selection comes down to a clear order of operations, and once you internalize it, the questions that used to feel tricky become fairly predictable. Start with prefix length and administrative distance, then route type, then cost, then ties. Back your answers up with the database and the routing table. If you want to go deeper, Cisco's official OSPF configuration guides and the relevant RFCs, such as RFC 2328, are the most reliable references to keep close while you study.
Comments
Post a Comment