Frame relay DLCIs

johnnynodough

I have read through at least 4 books on this subject and been all over cisco's site, and still have a conceptual roadblock. The DLCI is used to identify the path to the destination router in a frame relay network correct? The DLCI will map wither statically or dynamically (IARP) to an IP address. DLCIs are assigned by the ISP. Here are my questions.

Lets say my frame provide is AT&T. I have 3 sites, home office is assigned a global DLCI of 200. Two remote sites have been assigned global DLCI's of 201 and 202 respectively. When home office wants to chat with the first site of DLCI 201, it sends its packets to IP x.x.x.x which is mapped on the ISP side to DLCI 201, while in transit the DLCI gets swapped with 200. Reverse actions for site 1 to communicate to home office. Is this thinking correct?

Question 2. Same scenario, AT&T is my provider. I have a DLCI of 200, will there be any other router on the entire AT&T frame cloud that has the same DLCI? Is the importance of the DLCI only local? How does my home office no how to communicate through the correct PVC, or does it even care? Does the frame realy switch at the CO handle all that, knowing that this packet came in on circuit ID 74HCGS1111111PN, and wants to speak with ip x.x.x.x on DLCI 201, so it goes through this PVC on circuit ID 74HCGS999999pn?

I guess I can summarize my 2 questions by saying where does the importance of DLCI lye, locally or globally, can two companies in the same frame cloud have identical DLCI,s; and if so, how does the service provider know where to send data, with tables of circuit ID's?

If someone could summarize the life of a packet going from site A to site B, using IP address' and DLCIs, there importance, and even some thinking on the ISP side that would be great.

johnnynodough

damn, I must be the only person on the history of this board to not get this fully, or perhaps I am taking it to far. I understand that the DLCI is only locally significant since there is only 1024 possible DLCI values, and obviuosly there are a hell of a lot more than 1024 routers in all of AT&T's frame relay network. So I guess my question really is how does the frame provider no how to send data to the correct router since the DLCI is only locally significant? What is the purpose of a DLCI if it is only locally significant? How does it keep ABC company's site DLCI of 200 seperate from DEF company's site DLCI of 200? What about implementations of non serial links to frame such as ADSL to frame?

johnnynodough

Ok, I got an answer from out Cisco Channel Partner training coordinator. The answer is explained best in this scenario. Both companies exist in the same frame providers cloud

Company ABC - Data sent to main site using DLCI 200

Company DEF - Data sent to main site using DLCI 200

As the data hits the first DCE in the cloud, the customer is identified and the switch will arbitrarily switch DLCI's and append a proprietary unique identifier, usually a switch address, to keep data sent to the same DLCI's flowing to the correct endpoint.

That was my question, its above the CCNA level, but it was bugging me.

If anyone has anything to add please post. I would like to know a little more about what happens in the frame cloud from someone who knows, say if someone has worked at an ISP.

johnnynodough

And some further elaboration from one of the CCIE's in my Company, this relates to the uniqueness of common DLCIs on a frame cloud. Just passing along the knowledge.

Frames from different customers can have the same DLCI. With the 10 bit field, you can have a maximum of 1024 DLCIs per physical circuit. The combination of a physical interface/port and the DLCI ensures that the circuit is unique. Here is some sample cfg of a cisco router acting as a frame switch. The frame-relay route commands are doing layer two routing. They tell the frame switch "Hey, when you see DLCI XYZ on interface PDQ, send it to interface IOU and make it DLCI DFG".


bb1#sh frame route
Input Intf Input Dlci Output Intf Output Dlci Status
Serial0/0 102 Serial0/1 201 inactive
Serial0/1 201 Serial0/0 102 inactive

bb1#sh run int s0/0
Building configuration...
Current configuration : 189 bytes
!
interface Serial0/0
description Connected to R1
no ip address
encapsulation frame-relay
clockrate 64000
frame-relay intf-type dce
frame-relay route 102 interface Serial0/1 201
end
bb1#sh run int s0/1
Building configuration...
Current configuration : 236 bytes
!
interface Serial0/1
description Connected to R2
no ip address
encapsulation frame-relay
clockrate 64000
frame-relay intf-type dce
frame-relay route 201 interface Serial0/0 102
end

In this config, each physical interface (circuit) has one DLCI. S0/0 is announcing dlci 201 and s0/1 is announcing dlci 102. Again, this is a sample frameswitch config -- no the near/far end router. In this config, I could have used the same DLCI number on both sides such as:

bb1#sh frame route
Input Intf Input Dlci Output Intf Output Dlci Status
Serial0/0 18 Serial0/1 18 inactive
Serial0/1 18 Serial0/0 18 inactive

bb1#sh run int s0/0
Building configuration...
Current configuration : 189 bytes
!
interface Serial0/0
description Connected to R1
no ip address
encapsulation frame-relay
clockrate 64000
frame-relay intf-type dce
frame-relay route 18 interface Serial0/1 18
end
bb1#sh run int s0/1
Building configuration...
Current configuration : 236 bytes
!
interface Serial0/1
description Connected to R2
no ip address
encapsulation frame-relay
clockrate 64000
frame-relay intf-type dce
frame-relay route 18 interface Serial0/0 18

end

The combination of a the physical interface and the DLCI is what does the magic and makes it work.

ziggy

Yup, you are taking this a little bit deep. But it is better to understand it more than just a little. As far as the paths in the Frame Relay Switch, it is all done by the telco and it just works (PFM)

I work for a telecom company as a data communications installer. One of my primary jobs is to install the telco portion of WANs for customers to connect their router to.

When I go to a customer's premise, I will attach all devices and make all connections required to give the cicuit end-to-end continuity.

Such devices would include a CSU/DSU. I would then use a DST2000 (which is a test set designed to test end-to-end connectivity, DLCI, and LMI). It is an incredible test set. It emmulates the router sending out Status Inquire Messages and in turn receive Status Update Messages from the Frame Relay Switch. (among many other tests)

Once I am set up and all connections are made, I would then contact my peers many miles away who's job it is to make all Frame Relay path configuring in the Frame Relay Switch.

As far as the significance of Global/Local, I can not really answer. Other than I have installed close to a hundred Frame Relay circuits over the years and I have never really had to worry about the DLCI connections that are made in the Telco Frame Relay Switch.

The connections that occur within the paths in the Telco Frame Relay Switch are rarely/never a cause of problems and shouldn't concern the CCNA person.

A far greater problem that arises are conflicting LMI types between the Telco and the Router config.

johnnynodough

Yea, I knew I was taking it to far, I started realizing that once I couldnt find my answers in 5 different books, CCO, and googling like a mad fool

The truly entertaining thing is that Frame Relay is so painfully easy to configure, it was some of the logic behind it got me overly analytical. Usually its the config that is the tediuos part for me. I can configure it, and it works, but I wanted to know how it works end to end, and once I start I cant stop.

The other thing that got me was that packet switching networks are so incredibly fascinating to me, when you really sit and think about it, it is simply amazing how efficiently it works. Its like the equivalent of 100s of companies being connected to one LAN, utilizing one central companies devices, and having all the data be seperated and delivered to the correct destination. When you think of it in that analogy, of course you would be more curiuos of how your data travels through the ethernet

ziggy

Too funny, I am the same way at work......I need to know how things work.

But, I realized on the topic of how the Frame Relay "Stuff" travels through the "Cloud" was way too deep for me. I do talk to the guys that do the programming on a regular basis when I do the work orders.

Funny thing again....they do not even know how it works. I actually know more about Frame then they do. All they know is how to build paths using a PC and a simple GUI to the Switch. You could probably show a 14 year old that plays Playstation how to program a Frame Relay Switch using a GUI.

There is a good reason it is called the "CLOUD". Nobody really knows how it works.

johnnynodough

We just have to continue to believe the phone company


I think I am instituting a new rule in my training, if a topic takes 2 CCIE's to get me close enough to an answer to realize that I wont ever fully know, dont need to fully know, and really shouldnt fully know, I need to just abandon it, I'll add that to my personal goals list

Sorry, i just really wanted to use the smurf emiticon.

ziggy

QUIX0TIC

LOL