WAN – Frame Relay

Hello dear readers,

   In the last networking article we’ve talked about the role and functionality of ACLs. In this post I will focus on explaining another wide used WAN technology, Frame Relay. This is a WAN protocol that functions over the last two layers of the OSI stack, the Data Link and the Physical Layer. This protocol uses the concept of Virtual Circuits (VC) which is an identifier of the path that frames will travel through from source to destination. Frame Relay is often used to interconnect LANs over a WAN link. We will see how Frame Relay implements VCs that can change with every frame received, how to configure a router with Frame Relay, how a Frame Relay switch works so that in the end you will know the main concepts of this WAN protocol.
   Frame Relay introduced the concept of Virtual Circuits which unlike permanent leased lines offer flexibility and cost effective implementations. This WAN protocol is pretty simple to configure and also uses less equipment than other WAN implementations. It is cost effective because clients are paying only for the local loop lines and the dedicated bandwidth received from the service provider. In Frame Relay each virtual circuit is uniquely identified by a DLCI or Data Link Connection Identifier (DLCI) number. DLCIs are used to set the virtual path that a packet will travel to reach its destination. A disadvantage of Frame Relay is that this protocol does not support error detection mechanisms. When an error is detected, Frame Relay will simply drop the packet without notifying the sender or the receiver. We will see later that Frame Relay maps DLCI numbers to IP addresses. In every WAN transmission, DTE and DCE equipment must be installed between the two transmitting nodes. Frame Relay specifies how information is sent between these nodes (DTE-DCE) but does not specify how frames are moved between DCEs.
Frame Relay operates by using two main elements: the Physical layer and the link layer. The Physical layer is responsible for determining the electrical and mechanical specifications that must be used during transmission. The link layer specifies the protocol that will establish the WAN connection between the DTE and DCE. Let’s say information is sent from one node (DTE device such as a router) to another node in a remote location. The DTE will sent the packets to the closest DCE device, in this case it would be a WAN edge equipment such as a Frame Relay switch. Once the packet reaches the edge switch through the local loop, the client responsibility ends. How packets are sent between switches in the Frame Relay network is the ISP’s responsibility. In the end, the closest switch to the destination network delivers the packets to the client’s DTE device.
As I’ve told you earlier, Frame Relay uses the concept of Virtual Circuits to identify the logical path used to forward packets between two nodes. There are two types of VC that can be established:
Switched Virtual Circuits (SVC) – can change their configuration dynamically according to network changes.
   Permanent Virtual Circuits (PVC) – are configured by the carrier before any transmission can be made.
DLCI values are set by the the Frame Relay provider and have local significance only. This means that two DTE devices can use different DLCI numbers when sending data between them. DLCI numbers can be configured from 16 to 1007 while 0 to 15 and 1008 to 1023 are reserved. Another feature of Frame Relay is that a client’s DTE device can use multiple DLCIs when sending data to different destinations. Let’s take the following example: suppose we have three DTE devices (routers) A, B and C. A uses DLCI 100 for sending data to router B and DLCI 101 when sending data to router C. B uses DLCI 105 for sending packets to router A and 106 for sending data to router C. Remember that these numbers have local significance only. By using multiple DLCIs the cost is significantly reduced since the same physical devices are used.
Frame Relay receives packets from the network layer, encapsulates them into frames by adding DLCI numbers and checksums (CRC). Each frame is delimited by the 01111110 flag and then it is sent to the Physical layer for final delivery. Usually, Frame Relay topologies can be full mesh, partial mesh, star or hub and spoke. We have talked about these kind of topologies in the networking fundamentals articles. Frame Relay DLCIs are mapped to remote IP addresses. A DLCI would be used to forward packets to a certain network. In Frame Relay networks, inverse ARP is used to obtain the IP address (layer 3) of a remote network from the DLCI number (layer 2). Inverse ARP is enabled by default on all Cisco devices.  Remember that Frame Relay can support multiple protocols like IP, AppleTalk or IPX. The address mapping can be done in two ways:
dynamic mapping – a router will sent inverse ARP requests throughout the PVC to obtain the IP address for each hop. The router will then use the responses received to populate a local address table (also known as mapping table) that will be used for sending and receiving data.
static mapping – as a Network Administrator, you can configure static mappings between DLCI numbers and IP addresses. If you choose to assign a static mapping to an IP address, the dynamic mapping obtained by the inverse ARP protocol will be ignored.
Another aspect that you will need to remember about Frame Relay is that LMI (Local Management Interface) messages are exchanged between the DTE and the DCE equipment, to check the status of the Frame Relay connection. You can view the status by typing the show frame-relay lmi command from the privilege mode. By default the interval in which lmi messages are exchanged is 10 seconds. This interval can be modified using the keepalive command. There are many other aspects of the lmi mechanism, but are not needed for the CCNA exam. Feel free to add anything you know about lmi or Frame Relay in general, in the comments section.
We will continue talking about Frame Relay configuration commands. Given the following topology we will configure Frame-Relay on the these Cisco routers:

Frame-Relay topology

First, we’ll have to enable Frame-Relay on an interface, I will enable Frame-Relay on the interface serial 0/1/0 of router (R1):

Frame-Relay configuration
As you can see, I’ve added the IP address to the Serial 0/1/0 interface. I’ve then set the encapsulation type to frame-relay and finally configured the bandwidth used by the link (kb/s). Now I will do the same for router R2:
Frame-Relay commands
To verify your configuration use the show running-config or show interfaces serial [number] commands:
show interface serial command
We have just enabled Frame-Relay on these interfaces, our configuration will use dynamic mapping (with the help of inverse ARP). To configure static mappings on a Cisco device use the frame-relay map ip [ip address] [DLCI number] [broadcast] command. The broadcast parameter is optional and is used to enable broadcasts in a Frame-Relay topology. Remember that Frame Relay by default is a nonbroadcast multiaccess network (NBMA) and will not enable the transfer of broadcast or multicast traffic. Here is how a static Frame-Relay mapping would look like:
To verify our configuration, use the show frame-relay map command:
show frame-relay map command
   One big problem introduced in NBMA networks is caused by split horizon. Remember that this is a mechanism used to prevent routing loops by not sending messages to the interface from which the original routing information came from. This is a useful technique in routing protocols, but in Frame Relay configurations can cause problems since some of the Frame-Relay information could be blocked. The best way to resolve this is by using subinterfaces. A subinterface is a logical interface assigned to a physical one (a physical interface can support multiple subinterfaces). By using this technique you reduce the overall cost and also broadcast traffic can be forwarded between subinterfaces. These logical interfaces can be configured in two ways:
point-to-point – a Virtual Circuit is established between two subinterfaces (or between a subinterface and a physical interface). Each subinterface uses a different subnet IP address so that split horizon can be avoided. Each PVC has its own DLCI number and packets will be forwarded between subinterfaces using that particular DLCI
multipoint – a subinterface establishes multiple Virtual Circuits with one or more physical/logical interfaces. Interfaces that use the multipoint mechanism must be part of the same subnet.
When configuring subinterfaces, the physical interface must have the frame-relay encapsulation type configured first. Subinterfaces must be configured independently (with an IP address and mask that are part of a different subnet). I will show you how to configure Frame-Relay subinterfaces later.
   For your exam, you will need to know the elements that are needed to consider in a Frame-Relay implementation from the client’s side. As you already know, in this WAN protocol, the client leases only the connection between his DTE device and the ISP’s DCE. The aspects that the client must consider are: link speed (also known as access rate/port speed) and CIR (committed information rate). CIR refers to the guaranteed speed that the client can use when transferring information over the PVC (the ISP will offer a speed indicated by the CIR). A cool feature of Frame-Relay is that it offers support for speed bursting. This means that Frame-Relay can take advantage of the unused speed of an PVC. Because sometimes one PVC can have a higher usage than other, Frame-Relay can transfer the unused speed to the needed virtual circuit. An element called the CBIR (Committed Burst Information Rate) identifies the maximum speed that a link can support over the CIR. If a link has a CIR of 64 kb/s and a CBIR of 48 kb/s, this means that this PVC can use a maximum of 112 kb/s. In this case, if frames are sent with 112 kb/s speed, they are marked as Discard Eligible(DE) meaning that if there is not enough speed available, the frames will be dropped. Another element called the BE (Excess Burst) is used to indicate the remaining bandwidth of the access port. In our example, if the link can support a maximum bandwidth of 128, the BE is 128-112=16 kb/s.
   Two elements are used by Frame-Relay to notify devices about network congestion:
FECN (Forward Explicit Congestion Notification) – is a flag that signalizes the receiving DTE device that the link encountered congestion during transmission. Remember that frames with FECN flags set to 1, are sent only to the upstream devices. (devices through which frames will be sent to reach their destination)
BECN (Backward Explicit Congestion Notification) – a notification mechanism that informs devices, from which frames originated, that the PVC suffers from congestion. Frames with BECN flag set to 1 will be sent only to the downstream devices.
   At last, we will talk about configuring Frame-Relay subinterfaces on a Cisco device. First, we will need to enable the Frame-Relay encapsulation mode on the physical interface. To do this type the following:
encapsulation frame-relay
OK folks, now let’s configure one subinterface for the DLCI 100 and another one for DLCI 101:
The subinterface number can be chosen from 1 to 4294967293, I use the same number as DLCI for better identification. The frame-relay interface dlci [number] statement assigns the desired DLCI number for the interface. In this configuration, DLCI 100 will be mapped for the IP. There can be two options selected: point-to-point or multipoint. Remember that the multipoint option can be used when the same subnet is used by all routers in the Frame-Relay network. Usually, point-to-point connections use the /32 network mask, I just showed you an example. Now let’s do the same thing for the DLCI 101:
Frame-Relay configuration
It is important to remove any IP address configured on the physical interface!. If any IP was configured previously, frames will not be forwarded to the subinterfaces. You can verify/troubleshoot your Frame-Relay configurations by using the following commands:
show frame-relay lmi – view the status of lmi
You can even enable the lmi debugging mode to view the lmi exchanges in real time:
Debug Frame-Relay
show interfaces serial [number] – check interface configuration
show frame-relay map – verify inverse ARP operation (use the clear frame-relay inarp to clear the mappings table)
show running-config – view the running configuration
show frame-relay pvc [number] – checks the status of a PVC
I think that’s about it for this post, almost every part of the Frame-Relay protocol was described (at least I hope so). Please share to others and also post any comment/question. I hope you will enjoy this, stay tuned for the following WAN article. Have a wonderful day folks!

One thought on “WAN – Frame Relay

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s