Virtual Circuits

altjx

Ok, so according to the text, the frame relay switch switches the DLCIs in the frame relay header as it reaches its destination router.

What's confusing to me is for #4 in the "Did I Already Know This?", it says the correct answer is C. I thought the correct answer would be D. If R1 receives a frame with the DLCI of 222, wouldn't that mean that 222 is the DLCI on the sending router? Why is C the answer and not D?

To me, it looks like the text is contradicting, so can anyone clear this up for me? Thanks.

SharkDiver

C is the correct answer.

As in the middle picture, you need to think of the virtual pipe between router A and router B as being called "41" from router A's point of view, and being called "40" from router B's point of view.

Router A sends a frame to router B with a DLCI of 41. When the frame gets to the switch in the cloud that router A is attached to, and gets forwarded to the switch connected to router B, the DLCI is changed to 40. Router B knows that the frame came from router A because it came in with a DLCI that it knows is the pipe to router A.

DLCI is a local value, so right away you know that if R1 receives a frame with a DLCI of 222 in it, that it refers to a virtual circuit on R1.
222 is the DLCI of the pipe that R1 received the frame on, but only from R1's point of view.

altjx

SharkDiver wrote: »

C is the correct answer.

As in the middle picture, you need to think of the virtual pipe between router A and router B as being called "41" from router A's point of view, and being called "40" from router B's point of view.

Router A sends a frame to router B with a DLCI of 41. When the frame gets to the switch in the cloud that router A is attached to, and gets forwarded to the switch connected to router B, the DLCI is changed to 40. Router B knows that the frame came from router A because it came in with a DLCI that it knows is the pipe to router A.

DLCI is a local value, so right away you know that if R1 receives a frame with a DLCI of 222 in it, that it refers to a virtual circuit on R1.
222 is the DLCI of the pipe that R1 received the frame on.

Thanks for your reply..

I'm still a little confused. As in the picture, when Router B receives the frame, it has the local DLCI of the opposite router. But according to the question, which is separate from the diagram, R1 receives a frame with the DLCI of 222, why would that not be the local DLCI of the sending router (R2)?

In other words, in the picture above, when Router B receives the frame, wouldn't the DLCI of 40 be the local DLCI of Router B?

altjx

Ok, I think I'm getting a better understanding of this. The local DLCI is basically the DLCI of the opposite router.

So since R1 receives a DLCI of 222, it's basically saying that 222 is the DLCI from R1's perspective, so it's the local DLCI to that router.

I got it man... Damn that was confusing.

Forsaken_GA

The part you're missing is the fact that the frame relay switch swaps DLCI's, much like a router swaps MAC addresses. So when the FRS gets the frame in from R2, it has R2's DLCI. The FRS consults it's map and finds what circuit that DLCI corresponds to. So it says 'I received a frame on this circuit with the DLCI xxx, that's mapped to this outgoing interface that has DLCI yyyy'. It generates the new frame with the destination routers DLCI.

As far as why it's important for the DLCI to be that of the router receiving the traffic, that becomes obvious if you think about it. The DLCI is the frame relay equivalent of a MAC address. Unfortunately the frame relay frame format only has one identifier, unlike Ethernet, which gives you space for a Destination MAC as well as a source MAC. So if you only have one identifier for the traffic, it needs to be that of the destination, otherwise the destination wouldn't know the traffic is meant for it.

It's also important because these are virtual circuits. You can have a number of virtual circuits setup on one physical circuit. So if I've got a frame relay link that has 10 virtual circuits, and I receive traffic over my physical interface, I need to know which virtual circuit that traffic applies to, and that's what the DLCI does. The source, in this situation, doesn't work, because the same source could want to talk to hosts on many of my virtual circuits. So let's say the source wants to talk to hosts on 4 of my 10 virtual circuits. If the frame was tagged with the source DLCI, how is the destination router to know which of those circuits the traffic applies to? It can't, and that's why the DLCI needs to be a destination, not a source.

altjx

Forsaken_GA wrote: »

The part you're missing is the fact that the frame relay switch swaps DLCI's, much like a router swaps MAC addresses. So when the FRS gets the frame in from R2, it has R2's DLCI. The FRS consults it's map and finds what circuit that DLCI corresponds to. So it says 'I received a frame on this circuit with the DLCI xxx, that's mapped to this outgoing interface that has DLCI yyyy'. It generates the new frame with the destination routers DLCI.

As far as why it's important for the DLCI to be that of the router receiving the traffic, that becomes obvious if you think about it. The DLCI is the frame relay equivalent of a MAC address. Unfortunately the frame relay frame format only has one identifier, unlike Ethernet, which gives you space for a Destination MAC as well as a source MAC. So if you only have one identifier for the traffic, it needs to be that of the destination, otherwise the destination wouldn't know the traffic is meant for it.

It's also important because these are virtual circuits. You can have a number of virtual circuits setup on one physical circuit. So if I've got a frame relay link that has 10 virtual circuits, and I receive traffic over my physical interface, I need to know which virtual circuit that traffic applies to, and that's what the DLCI does. The source, in this situation, doesn't work, because the same source could want to talk to hosts on many of my virtual circuits. So let's say the source wants to talk to hosts on 4 of my 10 virtual circuits. If the frame was tagged with the source DLCI, how is the destination router to know which of those circuits the traffic applies to? It can't, and that's why the DLCI needs to be a destination, not a source.

Thank you, but I understand the entire concepts of how DLCI works. I was only confused because of the way the terms were worded and how the diagram was drawn out. I've completely understood where I was wrong.

Thanks for your additional explanation though.