RIP VERSION 1 AND 2

[joyride_us]

Hi,

if I start RIP version 1 and 2 on a router, will I have 2 routing tables built or 1 common table?

thank you

[thisisshanky]

Routing table is common. You cannot run both RIP V1 and V2 at the same time. If you enable RIP by default version 1 and 2 updates are received, but only version 1 packets are sent..(show ip protocol)

If "version 2" command is enabled the router sends and receives only version 2 packets.

(show ip protocol)

[Richard Burts · Accepted solution]

Sankar is correct that there is only a single routing table which would be fed by either or both of RIP versions 1 and 2. However, I disagree with Sankar on the question of running both on the router. In fact it is possible to run both RIPv1 and RIPv2 at the same time.

Sankar is correct that if you enter the version 2 command under router rip then the router will run only version 2. And if you enter the version 1 command under router rip then the router will run only version 1. But if you do not enter any version command, then the router will run both.

In particular consider these sections from the configuration guide for configuring RIP:

To control which RIP version an interface sends, use the following commands in interface configuration mode, as needed:

Router(config-if)# ip rip send version 1 2

Configures an interface to send RIP Version 1 and Version 2 packets.

Router(config-if)# ip rip receive version 1 2

Configures an interface to accept either RIP Version 1 or 2 packets.

I think that these make it clear that the router by default does run both versions 1 and 2.

HTH

Rick

[thisisshanky]

Rick is correct that you can configure the individual interfaces of your router to receive/send version 1, 2 packets to enable Rip V1 and V2 on the same router.

Quote: "But if you do not enter any version command, then the router will run both."

By default when RIP v1 is enabled, there is no version command that you put on the router. If the router was running both it should have send a broadcast packet (for RIPv1) as well as a multicast on 224.0.0.9 (for RIPv2). But it doesnt seem to do so...So I guess until you enable the "ip rip send/receive version" command to 1 2 both wont run.

Here is the setup:

10.1.1.1--lo--R1----172.16.1.1-(s0)-----FRSW----(s0)-172.16.1.2--R2

Here is a trace on R1 router. (debug ip rip)

00:09:52: RIP: sending v1 update to 255.255.255.255 via Loopback0 (10.1.1.1)

00:09:52: RIP: build update entries00:09:52: network 172.16.0.0 metric 1

00:09:52: RIP: sending v1 update to 255.255.255.255 via Serial0 (172.16.1.1)

00:09:52: RIP: build update entries

00:09:52: network 10.0.0.0 metric 1

00:09:52: subnet 172.16.1.0 metric 1 ip

I dont see a Mcast update being sent ...Any comments ??

[Richard Burts]

My first comment is a quote from your original post in this thread:

Routing table is common. You cannot run both RIP V1 and V2 at the same time. If you enable RIP by default version 1 and 2 updates are received, but only version 1 packets are sent..(show ip protocol)

The only way that the router can receive (and process) both kinds of updates is to be running both versions.

My second comment is that the ability to specify as an option sending or receiving both kinds of updates requires running both versions of the protocol. In your example the fact that a version 2 update is not being sent does not mean that version 2 is not running. It only means that the router has not been asked to send that version.

HTH

Rick

[Richard Burts]

I have thought some more about this discussion and I wonder if some of the discussion has been based on faulty assumptions. I wonder if some of us have been assuming that there was a version 1 process and a version 2 process and debating whether you could run both at the same time. I believe that there is in fact only a single process and that the Cisco code in that process does support both version 1 and version 2 of RIP. That process will listen for one or both types of updates depending on the configuration options chosen (by default it listens for both). That process will send one or both types of updates depending on the configuration options chosen (by default it sends version 1 but can be configured to send both).

A simple test could be set up to verify this. Set up 3 routers connected via Ethernet. Each router should have at least one other active interface (perhaps even a loopback interface) and should run RIP on both of the interfaces. Router A is the focus of the experiment and should be configured with the options to send both types of updates and listen for both types of updates. Router B should be configured with the version 1 command so that it only processes version 1 information. Router C should be configured with the version 2 command so that it only proceses version 2 information.

I believe that the result will be that router A will have in its routing table the other networks advertised by routers B and C which will demonstrate that it is running both versions. I also believe that router B will know the network from router A but not the network from router C. And that router C will know the network from router A but not the network from router B.

HTH

Rick

[joyride_us]

Thank you for your efforts. This is very valuable for me.

Another thing : although I chose to disactivate the auto-summary in RIPv1 (no auto-summary), it still sends the subnets summarized...is this normal ?

Thank you

[Richard Burts]

It is normal behavior that RIPv1 will always summarize network addresses at network boundaries. This is a basic behavior of RIPv1 and can not be changed. When you enter the no auto-summary command that will affect any version 2 updates but not version 1 updates.

HTH

Rick

[Kevin Dorrell]

The no-autosummary has no effect in RIPv1 - RIPv1 is so classful that it cannot help but to do autosummary on network boundaries.

The only time RIP will advertise a subnet is if the interface where it is being advertised is in the same major net and it has the same mask as the route to be advertised.

If it is being advertised out an interface on a different major net, then the major net will be advertised. That is, it will be auto-summarised. The router thinks it is a network boundary.

If the interface is on the same major net, but has a different mask to the route, then the route will not be advertised at all on that interface.

If the interface is on the same major net and has the same mask, but the mask is wrong at the other end of the link, then strange things happen. If the subnet is still on a subnet boundary according to the remote router's mask, then it will be interpreted by the remote router with the remote router's mask. If the remote router does not think it is a subnet boundary, then it will interpret it as a host route, /32.

The moral of the story is that RIPv1 can only handle contiguous consistent subnets, and will autosummarise as soon as you traverse a different major network.

Kevin Dorrell

Luxembourg