Frame relay : Replacing dedicated lines with frame relay

Frame relay
The original WANs consisted of multiple remote locations connected together. Each site may have a single computer or an entire LAN. The LANs connect to each other with dedicated lines provided by a telephone company. But the dedicated lines are expensive for customers, and they eat up too much of the carrier’s network capacity. Frame relay service is a solution to both of these problems.

Frame relay service uses variable-sized packets of data called frames. Unlike X.25, an earlier packet-switching service, frame relay service is a fast-packet technology. It discards erroneous packets, instead of correcting them. Error correction is performed at the end-points only, and not along the way, which results in a speedier transmission. If the receiving computer detects errors, it asks the sending computer to retransmit the data. Because errors are few with digital communication, this feature makes frame relay very quick. Unlike dedicated lines, frame relay is not a full-time connection. This is not a problem, because most businesses transmit data intermittently.

The PVC
Frame relay service providers set up a PVC between two customer sites that acts like a dedicated line. The customer chooses each PVC’s bandwidth. For example, a Louisville furniture manufacturer uses a 56-Kbps PVC to a remote facility in rural Kentucky but has a 256-Kbps PVC to the sales office in Atlanta, Georgia. More bandwidth is needed to the sales office, because more data is shared with this site. If the company used dedicated lines instead of frame relay, the cost would be almost double. Frame relay gives the best of both worlds: lots of bandwidth and low cost.

The CIR
The frame relay customer chooses the bandwidth of each PVC. As with other data services, the higher the bandwidth, the higher the cost. The 256-Kbps PVC costs more than the 56-Kbps PVC. The bandwidth is called the committed information rate (CIR), which is simply the rate of information that the phone company commits to always have available for you. The frame relay CIR is, therefore, the minimum speed limit. On the other hand, a 56-Kbps dedicated private line can transmit data no faster than 56 Kbps. The bandwidth of a dedicated line is, therefore, the maximum speed.

Frame relay is a “bursty” service. The furniture company could potentially transmit data at T-1 speeds across a 56-Kbps PVC if the phone company network has some spare bandwidth. Qwest boasts that its network has so much capacity that customers can save money by specifying “zero CIR” and still transmit data at T-1 speeds.

If frame relay service is used all within one LATA, then the service will be provided by the LEC in the area. For example, a Seattle hospital with numerous clinics in the same metropolitan area would purchase frame relay services from either its local carrier U S West or a competitive LEC that operates in the area. If the data network crosses LATA boundaries, a long-distance carrier, such as AT&T, will provide the service. However, the customer will still pay a local loop charge. The local loop is a dedicated private line from its facility to the long-distance carriers nearest frame relay-equipped central office. The local loop is provided by the LEC but will be billed on the frame relay carrier’s bill.

To install frame relay service, the customer must buy routers to be used at each location. Each carrier charges an installation fee, and the network technicians who program the router may charge additional fees. A business’ monthly frame relay pricing is based on PVCs, CIR, local loop charges, and any contractual discounts. Frame relay is a measured service—carriers show the usage on each invoice but most do not charge for it.

Replacing dedicated lines with frame relay
Consider the following example: A chain of tire stores in Ohio has four locations. Each location connects to the other three with dedicated T-1 lines. This “fully meshed” network requires a total of six T-1s. The charge for each T-1 includes the local loop at the starting point, the interexchange carrier mileage, and the local loop on the terminating end of the circuit. The network is illustrated in Figure 2.


Figure 2: Meshed network.


The company decided to replace its network of T-1 lines with frame relay service. Using frame relay, each location only requires one local loop connection to the frame relay provider’s central office. The data is then transmitted across the carrier’s network, which is usually called a cloud. To convert from dedicated lines to frame relay, the company had to purchase routing equipment and pay installation fees. The monthly charges are based on the local loop charges, the PVCs, and the CIR chosen by the customer. Figure 3 shows the change from dedicated lines to frame relay. The customer’s monthly cost dropped from $6,000 to $3,000.


Figure 3: Frame relay network.

Packet switching : What is a Packet?

Packet switching
Dedicated private lines tie up an entire phone line for the entire month. Circuit switching also ties up an entire phone line, but only for the duration of the call. Packet switching is much more efficient. Packet switching only ties up part of the phone line for the duration of the call. Like circuit switching, if the computers are not sending or receiving data, they stay off the phone lines, keeping them free for other users. But with packet switching technology, even the quiet times within one call will be filled with packets of data from someone else’s call.

The data is broken into packets that are sent from point A to point B. The packets may travel across different paths within the telephone company’s network. At the terminating point, all the packets are reassembled in the correct order.

What is a packet?
A packet is group of bytes of information that are processed independently across packet-switching networks (see Figure 1). Each packet has three parts: header, data, and footer. The header and footer are like a train with an engine in the front and a caboose in the back. They contain important information about the packet, such as the sender’s address, the destination address, the size of the packet, and the type of data contained in the packet, such as voice, data, or video. Voice and video packets are given priority in packet-switching networks, just like passenger trains are given priority on a railroad. Voice and video require a steady stream for the transmission to be smooth.


Figure 1: Packets, frames, and cells.


One reason packet switching is fast is because it does not correct transmission errors within the network. Instead, the data is sent again. The receiving computer tells the sending computer “I didn’t get all the packet … please resend the missing ones.” This is very effective for data communication where slight delays can be tolerated, but most businesses still refuse to move their voice traffic across a packet-switching network, because of the slight delay. Few companies want to sound like the 1980s computerman Max Headroom when they talk to their customers.

Prior to being transmitted across a computer network, a given computer file, such as a simple word processing document will be broken into packets of data. For example, imagine that a man named Sam in Seattle decides to write a love letter to Louise in Long Island. Sam plans to transmit the letter to Louise across a packet-switched network.

Using Microsoft Word, Sam writes a brief letter to Louise expressing his deep love for her. Sam is not a man of many words; his letter contains only 100 words. The size of the computer file containing the letter is only 20 Kb. Sam’s company uses a packet-switching service called ATM. The ATM service breaks Sam’s letter into 377 individual packets and transmits each one to the Long Island office. The whole process takes less than 1 second. In Long Island, the packets are reassembled in order, and the computer file is now accessible. Tears run down Louise face as she reads Sam’s love letter.

More?