Understand Frame Relay In 5 Minutes

CCNA12

The simplified easiest way to Understand Frame Relay:

1. the serial interface has no layer 2 address
2. the DLCI number is the ID of the virtual cable

Look at the diagram below:



99 is the ID for the virtual cable between R1 and the Cloud.
16 and 17 are the two virtual cables between R2 and the Cloud.
28 is the ID for the virtual cable between R3 and the Cloud.

The ISP cloud connects cable 17 to cable 28, cable 16 to cable 99.
Keep in mind that there is no Layer 2 address on the Frame Relay serial interface.
Ok, that is our physical connection.

Now , let's say we have IP address configured on these routers respective interface:

R1: 192.168.123.1
R2: 192.168.123.2
R3: 192.168.123.3

Now, we need to tell the routers how to reach others.

Router R2 uses cable 16 to reach R1, 17 to reach R3

R2:

Frame-relay map ip 192.168.123.1 16
Frame-relay map ip 192.168.123.3 17

Because of the ip 192.168.123.1 to the DLCI 16 mapping, packets destined for 192.168.123.1 will be put on cable 16.
Because cable 16 is connected to cable 99 by the Frame Relay ISP, the packets will go through cable 99 and arrive at R1.

Because of the ip 192.168.123.3 to the DLCI 17 mapping, packets destined for 192.168.123.3 will be put on cable 17.
Because cable 17 is connected to cable 28 by the Frame Relay ISP, the packets will go through cable 28 and arrive at R3.



Router R1 uses cable 99 to reach both R2 and R3, because that is the only cable connected to the cloud.

R1:

Frame-relay map ip 192.168.123.2 99
Frame-relay map ip 192.168.123.3 99

Because of the ip 192.168.123.2 to the DLCI 99 mapping, packets destined for 192.168.123.2 will be put on cable 99.
Because cable 99 is connected to cable 16 by the Frame Relay ISP, the packets will go through cable 16 and arrive at R2.

Because of the ip 192.168.123.3 to the DLCI 99 mapping, packets destined for 192.168.123.3 will be put on cable 99.
Because cable 99 is connected to cable 16 by the Frame Relay ISP, the packets will go through cable 16 and arrive at R2.
Now, R2 receives this packet which is destined for 192.168.123.3.
R2 checks its own ip to DLCI mapping, and would found 192.168.123.3 is mapped to 17.

R2 then put this packets on the virtual cable 17.
The packets would go through cable 17 and cable 28, and arrive at the right destination.




R3:

Frame-relay map ip 192.168.123.1 28
Frame-relay map ip 192.168.123.2 28

Because of the ip 192.168.123.2 to the DLCI 28 mapping, packets destined for 192.168.123.2 will be put on cable 28.
Because cable 28 is connected to cable 17 by the Frame Relay ISP, the packets will go through cable 17 and arrive at R2.

Because of the ip 192.168.123.1 to the DLCI 28 mapping, packets destined for 192.168.123.1 will be put on cable 28.
Because cable 28 is connected to cable 17 by the Frame Relay ISP, the packets will go through cable 17 and arrive at R2.

Now, R2 receives this packet which is destined for 192.168.123.1.
R2 checks its own ip to DLCI mapping, and would found 192.168.123.1 is mapped to 16.
R2 then put this packets on the virtual cable 16.
The packets would go through cable 16 and cable 99, and arrive at the right destination.

By David Rupu Xiao CCIE #24177

Forsaken_GA

but what if I want to use inverse arp? does that add another 2 minutes to my comprehension time?

tdean

i am confused about FR. everything i read refers to "frame relay switches"... arent they routers? also, where is it? i mean, if you work at company A, how do you manage your frame relay network? or do you just set up a router on your network that you want to communicate with the internet, and frame relay is part of the internet?

uugg.

chrisone

nice article and support tip for the community, you get green points just for that, but oh how i wish it were that easy to understand in 5 minutes hahaha

jamesleecoleman

The post was really helpful. Thanks!

DevilWAH

Well OK there's loads more to it yes, but that is a nice view of the fundamental idea.

I was told frame relay is like going to the air port. and the DLCI are the gate numbers. So if you are travelling to USA you go to gate A, if your going to the UK use gate B, India use C.. etc.... You don't need to worry about getting there, you will get on a plane and it worries about that.

Then at the other end you get off the plane and walk out a different gate at the destination airport. Now this gate could be a different number, or the same number. becasue the two airports are exclusive from each other. its only important I get on to the plane at the correct gate. every thing else after that will be up to the plane and air port operators.

All you are doing (when setting up the end points) is telling you packets what "gate" they need to start out on.

peanutnoggin

DevilWAH wrote: »

Well OK there's loads more to it yes, but that is a nice view of the fundamental idea.

I was told frame relay is like going to the air port. and the DLCI are the gate numbers. So if you are travelling to USA you go to gate A, if your going to the UK use gate B, India use C.. etc.... You don't need to worry about getting there, you will get on a plane and it worries about that.

Then at the other end you get off the plane and walk out a different gate at the destination airport. Now this gate could be a different number, or the same number. becasue the two airports are exclusive from each other. its only important I get on to the plane at the correct gate. every thing else after that will be up to the plane and air port operators.

All you are doing (when setting up the end points) is telling you packets what "gate" they need to start out on.

Excellent analogy... I will have to borrow that the next time I want to summarize Frame Relay! Thanks.

DevilWAH

peanutnoggin wrote: »

Excellent analogy... I will have to borrow that the next time I want to summarize Frame Relay! Thanks.

Psssssssssss.

You should probable give thanks to Jeremy from CBT nuggets I highly recommend his training videos. And he has loads of analogy's to help explain stuff.

Dilbert65

Frame relay usually are routers. The name "frame switch" is misleading. It is better to think of it as a "cloud" instead of switch.

I use a 2620xm as a frame relay cloud. If you notice the diagram the link between the routers is a cloud. If you wish to make a router a frame relay cloud you need a few serial ports.

One thing I wish to point out about the initial write up is that the commands are missing the "broadcast" on the "frame map" statements. Without those routing protocols wont run. Other than that a great post!!

tdean

isnt FR a NBMA? i thought that meant no broadcast?

EDIT: nevemind... i think i get that. however, im still struggling with the concept of OSPF over Frame Relay.

gouki2005

i understand static mapping but can you explain me inverse arp please...

tim100

tdean wrote: »

i am confused about FR. everything i read refers to "frame relay switches"... arent they routers? also, where is it? i mean, if you work at company A, how do you manage your frame relay network? or do you just set up a router on your network that you want to communicate with the internet, and frame relay is part of the internet?

uugg.

A Frame-Relay cloud consists of Frame Relay switches. Don't get confused between a lab environment and a true Frame-Relay cloud.

Frame-Relay is not part of the internet but rather a layer 2 packet switching technology. An example would be a Frame Relay connection between a customer router and an ISP router which are your endpoints. In between are Frame Relay switches which make up the cloud and have nothing to do with the internet. The internet connection is the connection between the customer router and the ISP router. Each endpoint is assigned a DLCI which makes up your PVC. The DLCI is locally significant and provides a way to identify the virtual connection. For example here is your virtual connection:

Router (DLCI 100) <

---

> ISP Router (DLCI 200)

Packets going out with DLCI 100 would eventually be packet switched to DLCI 200. The Frame Relay switches handle that part. So let's say you have 3 Frame Relay switches switching the DLCIs. A packet goes out DLCI 100 and this DLCI is the incoming DLCI on one Frame Relay switch. It will Switch that DLCI out to DLCI 101. The second Frame Relay switch sees DLCI 101 and switches it out to DLCI 201. The third Frame Relay switch sees DLCI 201 and switches it out to DLCI 200 and the packet finally arrives at the ISP router. The Router is the DTE and the Frame Relay switch is the DCE. The Frame Relay switch to Frame Relay switch connections are usually NNI (Network to Network Interface).

It would look like:

Router (DLCI 100) <

---

> FR1 (Incoming DLCI 100 out to DLCI 101) <

---

> FR2 (Incoming DLCI 101 out to DLCI 201) <

---

> FR3 (Incoming DLCI 201 out to DLCI 200) <

---

> ISP Router (DLCI 200)

Hope this helps.

tdean

Hi Tim, thanks for the response. that clears the process up for me. i understand the 2 end points. so the "middle" is literally many physical switches? and the virtual circuit is just a way to explain the end to end connection without really going into detail about each and every step (switches) through the cloud?

so, this is a NBMA b/c unlike ethernet, there are no broadcasts? it is packet switched the entire route though, and the route may differ each time, but end point would be the same....

are sub-interfaces always used with FR?

how does OSPF tie in with this? can they be used individually or must they be used in conjunction with each other?

tim100

tdean wrote: »

Hi Tim, thanks for the response. that clears the process up for me. i understand the 2 end points. so the "middle" is literally many physical switches? and the virtual circuit is just a way to explain the end to end connection without really going into detail about each and every step (switches) through the cloud?

Yes because the virtual circuit basically defines a logical point to point connection. You have terms like point to multipoint for topologies such as hub and spoke, partial mesh and full mesh but it's still basically a logical P2P for each pvc.

tdean wrote: »

so, this is a NBMA b/c unlike ethernet, there are no broadcasts?

This is where it gets a little tricky. Frame Relay is considered NBMA but if you configure the pvc(s) as broadcast as in a frame-relay map ip x.x.x.x [DLCI] broadcast statement or if inverse-arp is used then "copies" of a broadcast (or multicast) will be unicasted out each pvc. For example, if you configure OSPF on the router as network type broadcast then adjacencies will be formed and DR/BDR elections will take place as if it were a true broadcast network like ethernet. Caution must be exercised when configuring a Frame Relay network as broadcast as problems can arise. One example would be a Hub and Spoke topology. If a spoke if configured with more than one frame-relay map ip x.x.x.x [DLCI] broadcast statement then you will have redundant "broadcasts". If you configure OSPF in this topology you want to make sure the hub is the DR and the spokes are never DR or BDR.

tdean wrote: »

are sub-interfaces always used with FR?

No. If you don't configure a sub interface it is considered point to multipoint. If you do configure a sub interface you could specify point to point or point to multipoint.

tdean wrote: »

how does OSPF tie in with this? can they be used individually or must they be used in conjunction with each other?

OSPF doesn't tie into the equation. Frame Relay is layer 2 whereas OSPF is protocol number 89 at the transport layer.

Forsaken_GA

Dilbert65 wrote: »

Frame relay usually are routers. The name "frame switch" is misleading. It is better to think of it as a "cloud" instead of switch.

Frame relay is not a layer 3 technology, it is layer 2, so frame switch is a wholly appropriate moniker.

From your networks point of view though, yes, the frame relay network is one big cloud, you never know the details of the internals, nor is it necessary for you to know those details as long as the circuits are delivered properly

gmybft

The post was really helpful. Thanks!

doc holliday

Great post!

solnsusie

thanks loads, great explanation, but is this local DLCi, or global DLCI, I have no idea the difference between this 2, I need a good explanation on this.
thanks

mark076h

why did one of the routers have 2 cables in this example?

also what do yu do to configure the frame relay switch?

CCIEWANNABE

haha, i wish frame relay could be understood in 5 minutes..

and redistribution

and ipv6

and ospf

and every other ccie r&s lab exam topic...

till then, we all study on

phoeneous

mark076h wrote: »

why did one of the routers have 2 cables in this example?

also what do yu do to configure the frame relay switch?

Huh? Are you talking about the dotted lines? Those are the frame relay connections that the dlci's are mapped to.

In a lab environment you can configure a frame relay switch but in the real world the isp would handle that.

Jollycork

Personally, I had trouble with FR in the beginning because of the use of the "Clould" to signify connections between router DLCI circuits.

I had to actually setup my own "Clould" or Frame Relay switch with an NM-4A/S to 3 other routers to really get the idea down.

That and Chris Bryant's Frame Relay primer.

A picture of virtual circuits between routers through the frame relay switch showing the DLCI designations rather than just sticking the nebulous "Clould" might go a ways for the newbie. Did for me when I put it on paper.

Example: R1 has a DLCI 100 R2 has a DLCI 200 R3 has DLCI 300. each of those routers connects to one of the serial ports on the frame relay switch. How does the frame relay switch know where to send traffic and through what serial port? map the virtual circuits to each endpoint :

if DLCI 100 wants to go to DLCI 300, need a virtual circuit. If DLCI 300 wants to go to DLCI 100 need a virtual circuit for that as well. [have to have the return route]

Frame relay, least in my lab, will work without the ip address mapping [frame relay route] [no inverse arp] as long as there are virtual circuits to and fro .. run ripv2 on the routers and advertise the networks. I threw in a stub to the internet with NAT just to test everything...

Akiii

wow! I've just find Chris's lab, with every information. I think I'll put this together tomorrow..

JeanM

Really helpful post! Can someone post a link to Chris's lab?

MrXpert

Hi,
Although the OP's post is helpful it should really be mentioned that Frame relay cannot be simply understood in the 5 minutes it takes to read the initial post by CCNA12. Not mentioned is what type of interfaces are being used, physical,p2p or point to multipoint. It is clear this cannot be p2p because the frame map ip statements are being used but for someone new to FR, it should be explained. Also when setting up frame relay you really need to also work out whether you are going to be using OSPF, EIGRP and RIP and how this effects what you use. Other things to consider with it are lmi type,inverse arp and split horizon (assuming you need to use RIP or eigrp), is global or local DLCIs being used. There's a lot to consider and even for the CCNA you need to know about the different DLCIs as well as the different interface types.

2URGSE

tdean wrote: »

i am confused about FR. everything i read refers to "frame relay switches"... arent they routers? also, where is it? i mean, if you work at company A, how do you manage your frame relay network? or do you just set up a router on your network that you want to communicate with the internet, and frame relay is part of the internet?

uugg.

They are switches according to the reading I did.

There is not much management involved. You tell the ISP how much bandwidth you want and they give you the configuration for your router, you configure your routers and then test the connections. (of course you need some sort of routing protocol as well if you have multiple sites)

The frame-relay switch on the ISP's end is always the DCE and your router is the DTE.

Once the connections are up, there's not much more to it unless you run into issues or something goes down.