Why is route redistribution bad?

Question

Cisco

I often come across the statement that the redistribution of routes between different protocols is bad. No one will popularly explain why this is bad, or maybe the link will give "to read"?

Answer 1

Redistribution is bad because different routing protocols have different formulas for calculating the metric, and with redistribution you are figuratively "forcing" one protocol to rely on a route choice made by another based on completely different criteria. This is especially true for redistribution between link-state and distance-vector protocols. Yes, you can set up a metric during redistribution, but in the general case, all this only complicates the administration and troubleshooting of the network. There are many problems, you can “read” at least two of them at the link, a very reputable source. But I’ll repeat myself in this answer, a lot depends on the structure of your network and the point at which you want to redistribute, even for the same network everything can be from “very simple” to the real “hell” when debugging, in case the occurrence of problems.

Answer 2

Depending on where and where to redistribute. But mostly this is bad because of changes in prefix attributes when redistributing to another protocol - for example, metrics. Well, for example, I don’t think it’s worth explaining why the redistribution of fv BGP to igp is dangerous. But in simple cases everything works.

Answer 3

Thanks for the comments. Sori, but I can’t add to karma ...