Section 12.1. Overview of QoS

12.1. Overview of QoS

Communication networks face a variety of quality-of-service demands. The main motivation for a QoS unit in a data network port processor is to control access to available bandwidth and to regulate traffic. Traffic regulation traffic is always necessary in WANs in order to avoid congestion. A network must be designed to support both real-time and non-real -time applications. Voice and video transmissions over IP must be able to request a higher degree of assurance from the network. A network that can provide these various levels of services requires a more complex structure.

The provision of QoS to a network either does or does not come with a guarantee. Nonguaranteed QoS is typically based on the best-effort model , whereby a network provides no guarantees on the delivery of packets but makes its best effort to do so. In a non-real-time application, a network can use the retransmit strategy for successful data delivery. However, in a real-time application, such as voice or video networking, where timely delivery of packets is required, the application requires a low-latency communication. Consequently, the network must be able to handle packets of such applications more carefully .

Approaches to providing quality support can be further divided into integrated services or differentiated services . The details of these two approaches are discussed in Sections 12.2 and 12.3, respectively.