An autonomic PCN based admission control mechanism for video services in access networks

The introduction of new added value services such as IPTV has introduced great challenges for today's broadband DSL access networks as these services have stringent quality demands. In an attempt to protect the quality delivery of existing sessions, operators employ admission control mechanisms that limit the amount of sessions transmitted in the network. Current admission control mechanisms require a traffic specification of each stream, in order to know beforehand how many resources need to be reserved. For variable bit rate videos, which are bursty of nature, resources are reserved using the peak rate of the video. This leads to under-utilisation of the network as the amount of resources needed is over-dimensioned. We propose an autonomic measurement based admission control algorithm, optimised for the protection of video services in multimedia access networks. The algorithm is based on the IETF precongestion notification (PCN) mechanism and autonomically adjusts its parameters to the traffic characterisation of the video. The performance of this mechanism has been extensively evaluated in a packet based network simulation environment. Tests show that the autonomic nature of the algorithm leads to a better utilisation of the network while still avoiding any congestion in the network.