Questions on RIP

[RicN]

A few questions regarding the RIP protocols:

1. First, in your experience, is RIPv1/v2 used these days? If so, how common should you say that it is compared to static routing or OSPF?

2. Are there any security problems concerning RIP v1/2?

3. Does all RIP information has to fit into a single packet? If so, is there any limit on the number of routes?

4. I seem to remember a problem with RIP and variable subnet masks, is that still an issue?

[Matt Hobbs]

1. RIPv2 is what you would definitely use if you were going to go that way. The vast majority of networks I see are running OSPF, followed by basic static routes, and RIPv2 running last.

2. You can use MD5 authentication, can't remember if ProCurves support this though, quick look through the manual should verify this.

3. Good question, RIP can use auto-summary, so if you did have a large RIP network and you planned it correctly, it shouldn't be an issue. You'll run into other scalability problems before you hit that limit.

4. Not an issue with RIPv2.

Check this out, it's quite good: http://www.hp.com/rnd/training/technical/IRF.htm

[Matt Hobbs]

3. A single RIP packet can advertise 25 routes. It just sends more packets when there are more routes than this. However the maximum amount of routes is going to depend on the router you're using.

[RicN]

Matt Hobbs:

Thank you for that information!

>2. You can use MD5 authentication, can't
>remember if ProCurves support this though,
>quick look through the manual should verify this.

I did some looking, but could not find any command to enable this. Would that mean that the RIP "password" is just clear text and any device that could send RIP packets could be able to insert incorrect routing entries in the network?

[Matt Hobbs]

Looks to be the case, Most ProCurve devices supporting RFC 2543 (RIPv2) do not support RFC 2082 (RIP-2 MD5 Authentication).

Only the 8100 supports that.

Means that your only option for securing it is a clear-text password.

Another vote for OSPF...

[RicN]

Matt:

>Another vote for OSPF...

Does OSPF have better security options than RIP1/2?