2

I am learning about different dynamic routing protocols and I've learned that each dynamic routing protocol has an AD assigned to it which indicates the trustworthiness of the learned route from that protocol.

However, AD is only used if the routes to the same destination were learned through different routing protocols.

In what case would we use more than 1 dynamic routing protocol? Wouldn't it be enough to just use one, for example OSPF for the entire network, for example?

3
  • It is not just dynamic routing protocols. You get an AD of 0 for directly connected networks, and an AD of 1 for statically configured routes.
    – Ron Maupin
    Jun 15 at 19:29
  • Instead of "trustworthiness," you may want to think more along the lines of "believability." A directly connected network is completely believable, but an IGP with slow convergence is far less believable than one that converges almost instantly. Slow convergence can lead to routing loops.
    – Ron Maupin
    Jun 15 at 20:36
  • Just as an example, if I have a company with multiple locations, each with a dozen subnets that can’t be summarized into a single advertisement/announcement, connected to each other through MetroEthernet circuits, I may use EIGRP or OSPF to handle the routing between them to make it easier to manage changes on any of them, while then using BGP at each location to manage the peering with their respective ISPs for all internet traffic. I may even go so far as to have each site capable of using any other site for internet access as a redundancy.
    – Jesse P.
    Jun 18 at 12:26

3 Answers 3

7

The administrative distance isn't about "trustworthiness", it's about preference - a combination of cost, delay, bandwidth, resilience, loss rate.

External routing and internal routing generally differ in the trust you put in the sources, so you use BGP as EGP and an IGP like OSPF internally.

In a grown network there may be regions with different architectures, so you may have to integrate other routing protocols, at least for some time.

Trustworthiness (=absence of security threats) plays an important role in connecting different security zones by firewalls.

1

Keep it simple stupid. The KISS principle. You only want to use what is necessary to achieve the desired outcome, anything more is called overengineering which leads to downtime and security vulnerabilities.

Dynamic routing protocols are not always necessary even in multisite networks. In practice, you will sometimes see two dynamic routing protocols running concurrently when peering with other networks. This can happen when companies merge, or when there are multiple Internet links into a company, for example.

You want to avoid dynamic routing protocols if you can. Keep it simple stupid. There is a time and place to use them. Use wisely.

Finally, it's important to note that Administrative Distance takes a back seat to the most specific route. The most specific route wins everytime.

0

As others have mentioned, you would use an IGP(like ospf, eigrp, rip, isis) for internal routing within your organization and an EGP(the only one in use is BGP) for connecting between autonomous systems(like if you wanted to connect to a carrier PE for MPLS transport to another site). But there are other scenarios where you might use multiple routing protocols, especially when talking about provider networks. Like in the example of connecting to a provider network if you have two CEs that run eBGP to the provider, they could run iBGP between them.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.