In order to connect to a Frame Relay network, both DTE and DCE equipment needs to be located at the customer premises. This DTE equipment is usually a router, whose serial interface connects to a DCE device. In the past, customers required a completely separate DTE device known as a Frame Relay Access Device (FRAD) to connect to a Frame Relay network. However, almost all routers sold today (with an appropriate serial interface) are capable of handling Frame Relay encapsulation and communication. The DCE device is usually a CSU/DSU that provides clocking functions and the connection to the provider’s physical circuit. Ultimately, the physical link from the customer premises connects to the Frame Relay switching equipment of the service provider. This switching equipment is not the responsibility of the customer. The figure below illustrates the interconnections of equipment on a Frame Relay network.
Frame Relay is a packet-switching technology that exists at the Data Link layer of the OSI model and is one that has become increasingly common as a WAN solution since the early 1990s. Unlike with leased lines and circuit-switched networks, the available bandwidth on a provider’s Frame Relay network is shared amongst many subscribers. This sharing of resources leads to significantly lower costs than traditional leased lines.
Many people tend to be confused by packet-switching technologies like Frame Relay. Mostly this is a result of trying to understand how data actually gets from one location to another. On packet-switched networks (like Frame Relay), data streams are separated through the use of “virtual” rather than dedicated hardware circuits. In other words, a logical path is defined between endpoints, through a provider’s packet-switched network. Many virtual circuits will be defined for different customers, and will be multiplexed over the shared physical links of the network. As an example, consider the figure below. It shows two different companies, each connecting two offices over a provider’s Frame Relay network. Notice that between Frame Relay switches X and Y, both of their virtual circuits traverse a common physical link. The data that one company passes between their own offices is completely separate from the data of the other company – all data stays within each company’s dedicated virtual circuit only.
Two main types of virtual circuits can be defined on a Frame Relay networks – permanent virtual circuits (PVCs) and switched virtual circuits (SVCs). A PVC functions somewhat similar to a leased line, in that a service provider defines a path through the packet switched network to each customer location. In cases where companies wish to have “always-on” connectivity between locations using Frame Relay, PVCs are usually defined.
A switched virtual circuit (SVC) functions somewhat differently, almost like a circuit-switched connection. SVCs are not permanent, and can instead be created as required across a packet-switched network. For example, an SVC could be created between a company’s head office and a remote location. For the duration of the connection, data would travel across the path defined by the SVC. However, if the circuit was terminated and a new SVC established at a later time, data might travel over a completely different path.
Frame Relay networks are referred to as being non-broadcast multi-access (NBMA). What this means is that, by default, broadcast traffic will normally not be passed over a virtual circuit without explicit configuration. This is an important consideration when dealing with the use of broadcast-based routing protocols like RIP or IGRP in a Frame Relay environment. You’ll look at how broadcast traffic can be handled on Frame Relay networks later in this section.
A third and increasingly common WAN connectivity technique is known as packet switching. Unlike with leased lines, where customers pay for a dedicated link and consistent bandwidth, packet switching allows a service provider’s network resources to be shared amongst many customers, which in turn reduces costs. This makes packet switching a great choice for companies whose WAN traffic is variable or “bursty” in nature.
On a packet switching network, companies still connect to the provider network as they normally would, but instead of provisioning a dedicated circuit between locations, they share bandwidth will all other customers. The theory is that at any given time, a company will not be using its fully allocated bandwidth, based on the variable nature of data traffic. This allows other companies to make use of the available bandwidth, which in turn ensures makes more efficient use of the service provider’s network. Because the service provider doesn’t have to provision a physical end-to-end circuit for a packet switched customer, they are able to offer the service at a lower price.
A packet switched network provides a great example of the service provider “cloud” in action. When companies connect to the cloud, the service provider generally guarantees the minimum average bandwidth they will have access to, while allowing their traffic to “burst” to higher speeds if excess bandwidth is available on the shared network. On a packet switching network, individual packets are sent from one location into the “cloud”. These packets may take different paths to reach their destination, as per available bandwidth and network resources. When they arrive at their destination, they are reassembled in the correct order. In order to connect company offices, the service provider defines what are known as a “virtual circuits” between locations. Virtual circuits will be looked at in more detail later in the chapter. Common examples of packet switching WAN technologies include Frame Relay, X.25, and ATM.