Frame relay help? (partial mesh, full mesh etc)

Hi, having a little trouble understanding the different frame relay networks you can have. Please help/verify if you can. Thanks

Full mesh:  Each router is connected to each other. They are in the same subnet.  They address each other by using the dest IP with the dest DLCI
http://mvpbaseball.cc/full.jpg

Partial mesh:  Each router is NOT directly connected to each other. Rather, they all connect to one router with subinterfaces.  They are all differnet subnets. The main router does all of the routing in between them (is that right?)They address each other by using the dest IP with the dest DLCI
http://mvpbaseball.cc/partial.jpg


Hybrid mesh: A combination of full mesh and partial mesh.  Some of the routers are directly connected (via VCs ; multipoint), while the rest have point to point connections to a core router. Again, the core router does all of the routing in between the networks. They address each other by using the dest IP with the dest DLCI
http://mvpbaseball.cc/hybrid.jpg

And last but not least: Is the method of data delivery the same for each kind of mesh? In other words, say Router A (192.168.1.3  DLCI 13)   wants to send a frame to Router B ( 192.168.2.3  DLCI 23).  The way it sends the frame is the same no matter if partial mesh, full mesh or hybrid mesh is used right? The "frame switches" swap the DLCIs before delivering to destination? Sorry for the long post.
dissolvedAsked:
Who is Participating?
I wear a lot of hats...

"The solutions and answers provided on Experts Exchange have been extremely helpful to me over the last few years. I wear a lot of hats - Developer, Database Administrator, Help Desk, etc., so I know a lot of things but not a lot about one thing. Experts Exchange gives me answers from people who do know a lot about one thing, in a easy to use platform." -Todd S.

Dr-IPCommented:
It all depends on the data flow as to which one is optimal. For example, if you have several stores to connect up to the corporate office to send sales and inventory data, something quite common now days, a partial mesh solution, which I prefer to call hub and spoke as I think it better describes it, is a good solution. Since most of the data moves between the stores and the main office, and the loss of the hub which in this case is the main office, leaves the whole network as good as dead anyway, so there is no reason to go through the extra effort, and complexity to have a fully messed network.

Now let’s say we have a bank with several branches, and each one has their own server that keeps all that branches customers data locally. So that if they go to another branch to make a withdraw. The teller when they go to check their account for sufficient funds will be querying the server back at their home branch. So a hub and spoke solution would be inefficient since all date would have to pass through the hub even though it’s not its destination, and the loss of the hub would cripple the ability of the whole bank to conduct business. So a full mess solution is the best solution, since there is no central point of failure, is the most efficient, and the fact the operations can continue even if one of the branches network connection fails.

As for the hybrid solution, let’s go back to the first scenario, and add this to the mess. One of the stores has overgrown there location, which they don’t want to move from, so they decide to expand the store into the space goods where stored, and rent a warehouse down the street where customers would pick up goods. The way the operation works, is the customer makes the purchase in the store and pays for it, and a pick ticket is printed in the warehouse so by the time the customer arrives their goods are on the dock waiting from them. You could still use the hub and spoke solution, but if the main office goes down that store operations are badly impacted. So you mesh that store and its warehouse so if the main office goes down, the store can still communicate with its warehouse.



0
AutoSpongeCommented:
Those terms refer to frameswitch networks, not your routers.  Having a meshed network with routers means that each host is connected to at least two other hosts.  This defers from a hub-spoke or star configuration where one big router is connected to all the remotes.  Meshed networks are more stable and typically more efficient but implementation is often more expensive since you're paying for double the PVCs (or VCs if you're on ATM) to make sure every router has redundancies.
0
dissolvedAuthor Commented:
in a home/lab environment (using two 2501 routers and one 2522), what type of Frame Relay network could I have? Only partial right???  Would I be able to use the "map" command effectively?

Thanks. and thanks for the explanation
0
Ultimate Tool Kit for Technology Solution Provider

Broken down into practical pointers and step-by-step instructions, the IT Service Excellence Tool Kit delivers expert advice for technology solution providers. Get your free copy now.

Dr-IPCommented:
The best you could do is a quasi frame relay mesh at home, and that would be with the addition of a third router, and using the second serial interface on each to create a closed loop, but it still won’t be the same as a real frame relay mesh. Since you can only create that with an actual frame relay switch, something very few labs are going to have due to the expense and complexity involved.

As for the map command, you can use it on back to back routers, but there isn’t much to gain by it. Since you can only connect to one router per interface, which pretty much negates its usefulness.  

http://www.cisco.com/en/US/tech/tk713/tk237/technologies_configuration_example09186a0080094a3c.shtml
0
AutoSpongeCommented:
The map command is only really going to work for you when you have DLCI information coming from a frame switch.  Then you do show frame-relay pvc.  It will show either ACTIVE, INACTIVE, DELETED, or UNUSED.  If it's UNUSED, you have to add the DLCI in a map statement to the interface.  Then you'll see the PVC come active.  Back-to-back routers don't simulate this well, software might though.
0
dissolvedAuthor Commented:
Thanks guys. The thing I'm confused about is when you assign DLCIs.  Say I own a company with two sites. When I set up site A to communicate with Site B, do I need to know Site B's DLCI so I can map it to an IP ? (i'm assuming yes)

Also, take the following picture http://www.cisco.com/warp/public/125/14.gif

Spicey's config :
hostname Spicey
!
!
!
interface Ethernet0
 ip address 124.124.124.1 255.255.255.0
!
interface Serial0
 ip address 3.1.3.1 255.255.255.0
 encapsulation frame-relay
 frame-relay interface-dlci 140
!
------------------------------------------------------------------
Is this saying that Spicey's serial 0  has a DLCI of 140. So when other sites have to send traffic Spicey's way, they encapsulate with the DLCI of 140 in the frame relay header???

Thanks...
0
AutoSpongeCommented:
It means that he is known on the frame switch as <switch><card><port>140 and anyone on the remote end of his PVC will have to have their PVC built into the switch as <switch><card><port><near DLCI><switch><card><port>140.  Conversely, Spicey's PVC's will be built in the switch network as <switch><card><port>140<switch><card><port><remote DLCI>.

For instance, the old IBM switches had names like H005123.  Each switch has a shelf with cards.  Those cards are in slots, and each card as at least 1 port.  Everything before the DLCI is a physical assignment, the DLCI is logical.  So 1/2 of the PVC might look like H005123S2P3C140.  If Spicey talks to a router that is on DLCI 17 and connected to switch H005125 the full PVC might look like H005123S2P3C140H005125S1P2C070.  The only things the routers need to know is the DLCI because everything else is physical, and therefore out of their control.  So, to bind layer 2 DLCI's to a layer 3 IP address, you need a frame-map statement.  Layer 1 is taken care of.
0
dissolvedAuthor Commented:
So frames sent with a DLCIs with a 140, will be sent to Spicey

So, if  "Pracit"  wants to communicate with "Spicey", Pracit sends a frame to the frame relay switches with Spiceys DLCI of 140.  The frame relay switch receives it and sends it towards Spiceys way. The last switch swaps Spicey's DLCI (140) with Pracit's DLCI.  This is so the Spicey knows who sent the frame.

Is that accurate?

0
AutoSpongeCommented:
Yes. "The Frame Relay header contains the user-specified DLCI field, which is the destination address of the frame."
0
Don JohnstonInstructorCommented:
>Is this saying that Spicey's serial 0  has a DLCI of 140. So when other sites have to send traffic Spicey's way, they
>encapsulate with the DLCI of 140 in the frame relay header???
-or
>So frames sent with a DLCIs with a 140, will be sent to Spicey.

No.

DLCI's are the number that reference your end of a PVC. Think of it as an address (or label) of a circuit. If you want to send something and have it go over a paticular PVC (assuming you have more than one), then you will use the DLCI to tell the Frame-Relay Switch which PVC to send this frame on.

For example:

      Atlanta
      140  150
       |      |
   ----       ---
   |             |
 327          289
Boston      Chicago

If Atlanta wants to send data to Boston, it will put 140 in the DLCI field.
If Boston wants to send data to Atlanta, it will use a DLCI of 327.
If Atlanta wants to send data to Chicago, it will 150 in the DLCI field.

-Don
0
dissolvedAuthor Commented:
10-4
thanks guys
0
AutoSpongeCommented:
I misread the question slightly.  My statement is still correct, but the "destination" is the destination of the local switch, not the far end of the PVC.  In your scenario, Pracit sends his own DLCI to the switch to tell the switch which PVC he needs to forward the packet on.  As I stated earlier, the switch has the PVC built, so when he matches the first half, he immediately knows where the other end is and forwards the packets there.  In essense, the source defines the destination.  My router doesn't need to know the far end DLCI, the switch does that for me.
0
dissolvedAuthor Commented:
So a DLCI identifies a virtual connection between a DTE and DCE?


Atlanta
   |
   |  <---DLCI 140?
   |
F/R switch
   |
   |   <--DLCI 160?
   |
Boston


Boston want to communicate with Atlanta. Boston sends his IP combined with a DLCI of 160 to the F/R switch. The F/R switch looks at the DLCI of 160 and knows this goes to Atlanta....But how does it know this? That is the part I'm not sure on..
Thanks
0
Don JohnstonInstructorCommented:
> So a DLCI identifies a virtual connection between a DTE and DCE?

Correct.

>Boston want to communicate with Atlanta. Boston sends his IP combined with a DLCI of 160 to the F/R switch.

Almost. Boston puts Atlanta's IP address in the destination IP field and 160 in the DLCI.

The Frame Switch has a map (similar to an IP route) that tells it "when a frame comes in with a DLCI of 160, change the DLCI to 140 and switch it out port x". The frame then arrives in Atlanta with a DLCI of 140 having originated in Boston.

-Don
0
Don JohnstonInstructorCommented:
> So a DLCI identifies a virtual connection between a DTE and DCE?
Correct.

On second thought. No. I was a little too quick on the keyboard.

A DLCI identifies an end of a PVC.

-Don
0

Experts Exchange Solution brought to you by

Your issues matter to us.

Facing a tech roadblock? Get the help and guidance you need from experienced professionals who care. Ask your question anytime, anywhere, with no hassle.

Start your 7-day free trial
dissolvedAuthor Commented:
Thank you Dr IP, AutoSponge, and DonJohnston.
Much appreciated.
0
It's more than this solution.Get answers and train to solve all your tech problems - anytime, anywhere.Try it for free Edge Out The Competitionfor your dream job with proven skills and certifications.Get started today Stand Outas the employee with proven skills.Start learning today for free Move Your Career Forwardwith certification training in the latest technologies.Start your trial today
Routers

From novice to tech pro — start learning today.

Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.