Design of an Autonomic QoE Reasoner for Improving Access Network Performance

The access network has become an infrastructure for delivery of multimedia services, such as broadcast TV, voice over IP or video conferencing, to residential users . Compared to more traditional IP services such as e-mail, users demand a higher quality of experience (QoE) for these new services. To guarantee the quality of service (QoS), access network providers have deployed several service enablers, such as video on demand proxies. These techniques are only focused on the part between the service edge router and the access node, while the access network provider has only a limited view and impact on the home network environment. The QoE management in the home network is even more complicated by the wide range of user devices and the various home network configurations, containing wireless interfaces and home multimedia servers. This paper explores the architectural design of an autonomic intelligence component on the access node that determines required QoE enhancing actions, such as adjusting the video bit rate, by taking into account the specifics of the home network and the requested services. The viability of an analytical, fuzzy logic and neural network approach to implement an intelligent, autonomic reasoner is discussed, and experimental results are presented.