Background

QoS encompasses several technologies that enable selected applications to receive preferential access to shared resources, thus ensuring a specified level of performance for these applications. Pertinent performance parameters include jitter, throughput, latency, and packet loss. There are various approaches to allocating resources to attempt to ensure that these performance parameter values stay within acceptable ranges for specific applications or classes of applications.

When multiple applications are running simultaneously over the same network infrastructure, resources must be shared. When the Internet was first developed, connectivity was the primary objective. Most of the traffic was either file transfer or e-mail; the primary parameter by which to measure network performance was throughput.

With the rapidly growing role of multimedia and collaboration in current and future applications, connectivity alone is no longer adequate; emerging applications require interactive support from the network infrastructure. As a result, performance parameters such as packet delay, packet loss, and jitter are becoming just as important as throughput.

To support the diverse requirements of applications that will share tomorrow's network infrastructure, a portion of network resources must be allocated either to individual applications or to categories of applications.