Are there specific, concrete use cases for RIPv2 in modern network engineering?

Question

Due to its limitations (small hop count, slow convergence, possibly large periodic updates), RIPv2 has always been deemed inferior to alternatives like OSPF and IS-IS.

However, over the years, people used to argue that both configuration simplicity and low impact on CPU would remain somewhat useful for corner-case applications.

Does this still hold true? Nowdays, CPUs are faster, links are speedier, the other complex IGP alternatives are widely known.

Can anyone please point specific use cases for considering RIPv2 in modern networking engineering?

Answer 1

I have often deployed RIP on the edge of MPLS VPNs on PE-CE links to give customers dynamic routing into their own VRF to good effect (though admittedly less and less these days).

In this case it is a perfectly valid use - small RIB size (especially with a default out, local-site prefixes in), convergence time often isn't an issue (often only a single link to the site), hop count is fixed (eg: 1). Due to the meager control-plane requirements, this provides a very scalable solution for hundreds or thousands of CEs per PE.

As Ron has mentioned - not all CE devices support BGP, and no (sane) carrier wants to run OSPF or ISIS across a network boundary.

It seems to have become very hip in the last decade of Networking to hate on RIP or imply that using it is somehow a reflection of poor design/architecture decisions, but I tend to prefer simplicity and The Right Tool for the Job™.

Answer 2

RIP is nearly universally supported, even on small routers which don't support OSPF or IS-IS.

Suppose you had a network with only two routers with a few LANs connected to each one. It would be easier to set up RIP than it would be to configure static routes, and any changes to the networks on one or the other router would automatically propagate.

If your routers don't support OSPF, or its configuration is more complex than warranted for your site, RIP could work.