The RIP Routing Protocol

[dallas31]

Greetings,

I am not even sure if this is the correct place to ask about this but, can someone please help clarify something for me. At which layer of the OSI model does RIP work at? I always thought that it worked at the network layer, however, I am now seeing documentation that it works at the application layer because of its use of UDP.

Thanks,

Dallas

[stomasko]

Rip works at the network layer. All routing protocols are connectionless with the exception of BGP which uses TCP.

Hope that helps.

Steve

[dnewell24]

The RIP process does reside at the application layer because it uses UDP port 520 but its purposes are to maintain the Routing Information Base. The useful information or the job RIP performs affects the network/internet layer.

HTH

Ryan

[Richard Burts]

To ask what layer of the model a routing protocol is in is to some extent a misleading question. And trying to answer it based on whether it uses a transport protocol like TCP (for BGP) or UDP (for RIP) may lead to misleading conclusions. If you look for transport protocols you would wind up proving that EIGRP and OSPF which run as IP protocols are different from RIP and BGP.

It is probably accurate to say that all routing protocols are applications. All the routing protocols process input and make certain decisions. And the purpose of the routing application is to supply information which is used at layer 3 of the model.

HTH

Rick

[Kevin Dorrell]

Rick,

This question has come up before, and there are always as many different answers as people who want to answer it. I was thinking about why it is a misleading question. I think it is because the model is about data communication, i.e. the data plane. Routing protocols belong in the control plane - they are not part of the data path itself. That is why their position is ambiguous. It's like asking which layer is network management. What do you think?

Kevin Dorrell

Luxembourg

[Richard Burts]

Kevin

Yes this kind of question comes up from time to time. Given the importance that many of us attach to understanding the layered model it is certainly understandable that many people would question where various things fit.

I think that you have a good approach in analyzing it as a data plane versus control plane issue. I had not thought of it this way before but I think this approach works pretty well for this question.

HTH

Rick