End-to-End Quality of Service

This section introduces the notion of Quality of Service as it is currently provided by the Internet. As of today, the Internet offers a point-to-point delivery service, which is based on the "best effort" delivery model. In this model, data are delivered to their destinations as soon as possible, but with no bandwidth or latency guarantees. Using protocols, such as TCP, the highest guarantee the network provides is reliable data delivery. This is adequate for traditional applications such as FTP and Telnet, but inadequate for applications requiring timeliness guarantees. For example, distributed multimedia applications need to communicate in real-time and are sensitive to the quality of service they receive from the network. For these applications to perform adequately and be widely used, Quality of Service (QoS) must be quantified and managed. Up to now, QoS has been specified in terms of raw bandwidth and latency for network resources, or CPU and memory utilization for system resources, and the network infrastructures have been deployed to support real-time QoS and controlled end-to-end delays. However, at present the notion of QoS extends from the communication level up to the application level, in order to map QoS application requirements into low-level QoS parameters. One distinguishing feature of QoS is that the application-level QoS parameters are subject to negotiation between applications and both the system and network components, in order to determine whether the application requiring QoS can be provided by the entire system. Several resource managers (brokers) can be deployed to provide support for the dynamic management of QoS. In this chapter the traditional concept of Quality of Service at the network level is introduced, a taxonomy of the applications and their QoS requirements is provided, and a more recent notion of end-to-end QoS is presented.