QoS can be defined as the "measure of transmission quality and service availability of a network (or internetworks)."[1] Another definition of QoS is that it "refers to the ability of a network to provide improved service to selected network traffic over various underlying technologies."[2] The common theme here is that QoS ensures quality service to network traffic. Recall from Chapter 1, "Network Design," that QoS is an intelligent network servicea supporting, but necessary, service provided by the network. QoS is not an ultimate goal of a network; rather, it is a necessary service that enables network applications. In contrast, voice communication is an example of an intelligent network solutiona network-based application that requires the support of network services, including QoS. A network within which no QoS strategy, tools, or techniques have been implemented treats all traffic the same way and is said to be offering a best-effort serviceit does its best to send all packets and treats all packets equally. So, if a company's CEO is on a voice call with an important client and someone starts to download a movie to watch over the weekend, the network treats both types of traffic equally and does not consider voice traffic any differently if contention for network resources exists. This is probably not the way the CEO imagined the network should work. The QoS strategies presented in this chapter can be used to ensure, for example, that voice traffic takes priority over movie downloads. A converged network is one in which data, voice, and video traffic coexist on a single network. These diverse traffic types have different characteristics and hence different quality requirements. The QoS tools introduced in this chapter are designed to improve the QoS in networks with a variety of traffic types. Specifically, the QoS parameters affected are the factors that affect the quality of the service provided to the transmission of traffic: packet loss, delay, and jitter. Packets are typically lost because of network congestion. The effect of the loss depends on the application being used. For example, loss of a single voice packet is not detrimental to the quality of the voice signal at the receiving end because it can be interpolated from other voice samples; loss of multiple voice packets, though, can cause the received message to be unintelligible. On the other hand, a packet sent through the Transmission Control Protocol (TCP) (for example, a file sent with a File Transfer Protocol [FTP] application) that is lost would amplify the congestion problem because it would have to be resent and would therefore consume more bandwidth. Delay, also called latency, is the time it takes packets to travel through the network. Delay has two components: fixed and variable. These terms are described as follows:
Jitter is the variation in the delay experienced by packets in the network. In the example of jitter illustrated in Figure 6-1, the sender sends the data out at consistent time intervals, Dt. The receiver is seeing a variation in the delay of received packetssome are greater than Dt while others are less than Dt. Jitter is usually not noticeable for applications such as file transfers. However, applications such as voice are sensitive to differences in packet delaysfor example, a listener might hear silent pauses where none should exist. Figure 6-1. Jitter Is the Variation in the Delay of Received PacketsNote Special dejitter buffers are incorporated into voice-enabled routers to smooth out the differences in packet delays by converting the variable delay to a fixed delay. However, these dejitter buffers increase the overall delay in the network. QoS allows you to control and predict the service provided by your network for a variety of applications. Implementing QoS has many advantages, including the following:[3]
|