Understanding Frame Relay by doing with GNS3 - Part #1

Frame Relay is one of the most successful WAN technologies that ISPs used to have in their networks, today, many network engineers still need to learn it in order to be able to understand the newer technologies in modern networks. In this article, I am going to show you how to setup frame relay from both the ISP and the customer perspective. Let’s get started with the definition of some basic concepts you should know about frame relay.

1. Permanent Virtual Circuit (PVC)

This is a nonphysical Frame Relay communication link between our WAN devices. The users see the connectivity just like a physical link between their remote devices. In reality, this link is not physical and there is no actual bandwidth allocated to a clear cut physical connection between the devices. The PVC, is set up at both the customer and the ISP’s location at the moment the contract is signed and will stay for the period of the contract. With a Switched Virtual Circuit (SVC), is another way of setting up the links between devices, in this case, the circuit is set up when there is a desire to communicate, it is not permanent, just like a phone call.

1. Data Link Connection Identifier (DLCI)

The DLCI is like the MAC address in Ethernet networks, it is used to send communications on the link, but works quite differently as compared to the MAC address communications. When a device needs to send communication to another device, it just needs to know the DLCI to use to go out to reach that device, no need to know the distant DLCI, this is a locally significant number-another device on the cloud can use the same number-. To better understand how DLCIs work, take the analogy of gate numbers in airports, To flight from Mauritius to DUBAI, you need to go to gate A4, this is only locally significant to Mauritius. When you get on the aircraft and arrive to DUBAI, you may land at get B15. The only thing we needed to care about when we departed was gate A4, we don’t really care about the arrival gate except when we come back, we might be invited to use gate B15 at DUBAI Airport. DLCIs are represented using 10 bits, so there are 2^10, or 1024 different DLCI addresses possible. Keep in mind that some of these are reserved, however. For example, DLCI 0 is used for Local Management Interface (LMI) signaling. A common approach for Frame Relay providers is to start DLCI numbering at 100, and to count by 5 or 10.

1. Local Management Interface (LMI)

LMI are signalling packets exchanged between the Frme Relay Switch and the customer border device. They are used to send information about the status of the circuit, this information include DLCI configurations, Circuit statistics, and any problem that can occur.

1. Inverse ARP (InARP)

Inverse ARP is used by devices participating in the Frame Relay topology to resolve the layer 3 address of the remote peer knowing the local corresponding DLCI.

1. Committed Information Rate (CIR)

The CIR for a Frame Relay virtual circuit is the rate at which the provider agrees to accept data from the customer. Depending on the terms of the contract and the availability of ISP’s bandwidth, you could be allowed burst (send data) above this value at times. It is measured in Bps.

1. Local Access Rate (AR)

The data rate of the user access channel. The speed of the access channel determines how rapidly (maximum rate) the end user can inject data into a frame relay network. In the real world, the CIR will be in most cases considerably less as compared to the Local Access Rate, the way the ISPs make sure the customer do not overwhelm their network by this hight rate of data transmission is by using Frame Relay Traffic Shaping (FRTS) to ensure that customers behave and send traffic at the required rate.

1. Forward Explicit Congestion Notification (FECN)

The FECN is a technique used by the Service Provider to notify the customer to slow down in case saturation is noticed. The DCE equipment or Frame Relay Switch sets the FECN bit and send the frame forward in the cloud. This notifies the receiver that there was congestion experienced during the transmission.

1. Backward Explicit Congestion Notification (BECN)

BECN is the opposite of FECN, in this case, the DTE set’s the BECN bit to notify the Service provider that congestion avoidance should be initiated.See more at: 

• https://www.cisco.com/c/en/us/support/docs/wan/frame-relay/47202-87.html
• https://tools.ietf.org/html/rfc3133

Now lets take a look at some frame relay design considerations. We will describe here the various topologies and configurations mode you can have. 

1. Frame Relay topologies

In designing any WAN, regardless the switching type, there are three basic design approaches that can adopted : Star, Fully-Meshed or Partially-Meshed topologies. Frame Relay can be used in one of those three topologies. Frame Relay is also a nonbroadcast multi-access (NBMA) connection scheme.     

1. Star or Hub and Spokes topology: In this topology, remote sites are connected to a central site, which usually provides a service. Star topologies require the fewest PVCs, making them relatively inexpensive. The hub router provides a multipoint connection using a single interface to interconnect multiple PVCs.   

1. Fully-meshed topology: In a fully-meshed topology, all routers have virtual circuits to all other destinations. Although it is expensive, this method provides redundancy, because all sites are connected to all other sites. Fully-meshed networks become very expensive as the number of nodes increases. The number of links required in a fully-meshed topology with n nodes is (n – (n – 1)) / 2.
   
2. Partially-Meshed topology: In a partially-meshed topology, not all sites have direct access to all other sites. Connections usually depend on the traffic patterns within the network.

No matter the topology that you chose, you can decide to have a point-to-point or a multi-point-to-point configuration, the choice of one of those depends largely on the routing policy of the customer, in both situations, the ISP configuration stays the same. Let’s go now to the configuration.

For the configuration, we will consider the following network topology: