Frame Relay - question about the IP addresses

2URGSE

Sort of related to:

http://www.techexams.net/forums/ccna-ccent/75545-configuring-frame-relay.html

I'm doing a point to point Frame Relay lab.

R1 and R2 are on the following subnet: 10.10.10.0/30

Now, I did the reading on frame-relay and the most important thing that was mentioned was:

The DLCI is significant to the local router and the IP address is significant to the remote.


The diagram shows R1 S/0/0/0 with IP 10.10.10.1 and R2 S0/0/0 with IP 10.10.10.2

So far so good, I do realize that when you create a sub-interface, it's a logical interface that will recieve an IP address and there is no need for an IP address on S0/0/0.

So....

R1 configuration looks like this:

R1(config-if) encapsulation frame-relay
R1(config-if) int s0/0/0.1 point-to-point
R1(config-subif) frame-relay interface-dlci 102
R1(config-subif) ip address 10.10.10.1 255.255.255.252



I finished the lab by running

show frame-relay map
show frame-relay pvc

Everything looks good.

Am I not understanding the theory correctly? why are they mapping the local IP address of R1 and not R2? I thought the IP address mapping of the frame relay goes to the remote router, can someone please clarify.

2URGSE

Reading

http://www.techexams.net/forums/ccna-ccent/52527-understand-frame-relay-5-minutes.html

trying to make sense of the IP address.....

Rakurai

That isn't the mapping of a local IP address, it is just the IP address for the interface. Mapping an IP address for remote will look like:

frame-relay map ip IP Address DLCI (broadcast)

And the broadcast is only if necessary.

EDIT:

Here is a article on Ciscopress that addresses the frame-relay map command. It does say that the command is used to map ip address to multipoint subinterfaces, when you are not using inverse-arp.

http://www.ciscopress.com/articles/article.asp?p=170741&seqNum=6

2URGSE

Rakurai wrote: »

That isn't the mapping of a local IP address, it is just the IP address for the interface. Mapping an IP address for remote will look like:

frame-relay map ip IP Address DLCI (broadcast)

And the broadcast is only if necessary.

EDIT:

Here is a article on Ciscopress that addresses the frame-relay map command. It does say that the command is used to map ip address to multipoint subinterfaces, when you are not using inverse-arp.

Using Frame Relay Point-to-Point Subinterfaces > Cisco Frame Relay Configurations

I understand what the command does, I understand when to use it.

What I do not understand is, they say the IP address should be that of the remote router, but then they are using the local IP address. WHY?

boredgamelad

frame-relay map takes care of the L2 and L3 addressing in a single command, by saying "map this remote IP address to this DLCI". So frame-relay map ip 10.10.10.2 102 issued on R1 would tell R1 "the destination 10.10.10.2 can be found on DLCI 102".

The use of point-to-point subinterfaces assumes that the router on the other end of the PVC will be configured with the right IP and subnet mask (or, in Cisco's words: "For point-to-point subinterfaces, the destination is presumed to be known."). The frame-relay map command isn't used, so we don't need to tell the R1 what the IP is for R2.

R1(config-subif) frame-relay interface-dlci 102 // This command tells the router "for communication on subinterface s0/0/0.1, use DLCI 102". This takes care of the L2 addressing only.
R1(config-subif) ip address 10.10.10.1 255.255.255.252 // This command tells the router "the ip address of s0/0/0.1 (my local logical interface) is 10.10.10.1/30".

So if we do a show frame-relay map:

R1#show frame-relay map
Serial2/0.1 (up): point-to-point dlci, dlci 201, broadcast, status defined, active

and ​show route:

10.0.0.0/30 is subnetted, 1 subnets
C 10.10.10.0 is directly connected, Serial2/0.1

This is all the router needs to know what to do with a packet meant for 10.10.10.2. It will look in the routing table and forward it out S2/0.1, which is mapped (using the frame-relay interface-dlci command) to DLCI 201.

In summary: You don't use the IP address of the remote router when configuring point-to-point subinterfaces because that's just how the configuration works. None of the commands require the remote IP address, so you don't need to use it. All you have to do is tell the subinterface which DLCI to use and your router will assume the other end is set up correctly.

Forsaken_GA

The first mistake you're making is trying to figure out mapping on a point to point interface.

The initial example you're using involves no mapping. You tell the router what the subints ip address is, and what DLCI is mapped to it. And that's it. All traffic sent out that subint will be addressed with a source of the ip configured on the interface, and the dlci defined by the frame-relay int-dlci command.

I've lost track of how many times I've said this - On a point to point interface, you do *NOT* need to do any layer 2 to layer 3 mapping (that's what mapping the DLCI to an IP is doing, l2 to l3 mapping). By it's very definition traffic on a point to point link has only one place to go - the other side of the link. So in a frame relay point to point interface, all you need to do is tell it which DLCI it belongs to, and what it's IP is, and you're good to go.

On a multipoint interface, you need to map the remote IP to the local DLCI, because there could be more than one, so the router knows which DLCI to send the traffic out on (it's not technically correct to say the DLCI is only locally significant to the router... it's actually link local, the frame relay switch cares about the DLCI too, but that's transparent to the user on the router)

2URGSE

Forsaken_GA wrote: »

The first mistake you're making is trying to figure out mapping on a point to point interface.

The initial example you're using involves no mapping. You tell the router what the subints ip address is, and what DLCI is mapped to it. And that's it. All traffic sent out that subint will be addressed with a source of the ip configured on the interface, and the dlci defined by the frame-relay int-dlci command.

I've lost track of how many times I've said this - On a point to point interface, you do *NOT* need to do any layer 2 to layer 3 mapping (that's what mapping the DLCI to an IP is doing, l2 to l3 mapping). By it's very definition traffic on a point to point link has only one place to go - the other side of the link. So in a frame relay point to point interface, all you need to do is tell it which DLCI it belongs to, and what it's IP is, and you're good to go.

On a multipoint interface, you need to map the remote IP to the local DLCI, because there could be more than one, so the router knows which DLCI to send the traffic out on (it's not technically correct to say the DLCI is only locally significant to the router... it's actually link local, the frame relay switch cares about the DLCI too, but that's transparent to the user on the router)

I've got it now.

I have to say in my defense that most of the simulators don't explain things very well, if they would have just said that the IP address is the one assigned to the sub-interface, it would have been much more clear.

I've been in support for 15 years, people should not assume others know what they are talking about, they should give me details.

Anyways, I get it now.

Thanks.