Along with the commercialization of the internet, the distribution of new services such as VoIP (Voice-over-IP) or VPN (virtual private networks) as well as the use of IP networks even for "mission-critical" applications e.g. ERP (enterprise resource planning), there has also been a permanent increase in demand for bandwidth. Apart from this, increasing user expectations with regard to performance, security and service have to be met. To achieve these goals the services offered by the ISPs must be differentiated or ranked according to their varying requirements and then treated accordingly for transport. Therefore the important role of quality of service (QoS) in IP networks is receiving more and more attention from the research community.
The traditional structure of the internet as a "best-effort" network allows only rudimentary support with respect to the differentiation of services and has to be extended appropriately to enable QoS. The goal of QoS is to achieve predictable behavior for fundamentally important parameters (bandwidth, error rate, delay and jitter) even in overstressed network nodes.
Important functional parts and a basic formulation of questions for QoS are: