Ph.D. Thesis Resources optimization in multimedia communications

vi Foreword University of Cagliari vii Foreword Since my M.Sc. dissertation [SR02], I have focused my research interest in the Information and Communication Technology field, with regard on the communication network resources optimization. While in [SR02] I have analysed a single network component (a multi-service switch), in this Ph.D. dissertation I have enlarged my focus to entire infrastructures based on different technologies: fixed network (using multicast services), wireless and mobile networks. Nevertheless, the thread flowing through these topics is always the same: the optimization of the resources involved. In particular I have faced the following themes: − QoS (Quality of Service) optimization for multicast services. − Power optimization for Wi-Fi (Wireless Fidelity) stations. − 3G (Third Generation system) performance optimization. In the following, a brief description of my contribution in these fields. QoS optimization for multicast services. It focused on the group multicast routing problem: how to allocate network resources when several multicast sessions occur. The major contribution to this problem was twofold. Firstly, the use of the genetic algorithms (GAs) was proposed for solving the complexity problem inherent to the grouping of multiple sessions. Secondly, a novel cost function was proposed, weighting in a single expression both network bandwidth allocation and provided one-way delay. The proposed function was guided by few parameters that can be easily tuned during traffic engineering operations; an appropriate setting of these parameters allows the operator to configure the desired balance between network resource utilization and provided QoS (in terms of transmission delay). Experimental results were compared with the ones obtained by applying GA technique to two cost functions proposed in literature for solving this kind of problem. This highlighted that the proposed method was able to provide better solutions both in terms of link usage and end-to-end delay requirements. Power optimization for Wi-Fi stations. In Wi-Fi infrastructured LANs (Local Area Networks) with Distributed Coordination Function (DCF), the power saving algorithm works as follows: the stations with no frames to transmit may switch to the doze mode to save power; all these stations are informed by the Access Point (AP) at the beginning of each beacon interval if there are pending frames; these stations then start contending the channel to receive these pending frames from the AP, if any. In this dissertation, a different strategy is proposed, based on giving the AP the power of deciding which stations with pending frames to notify. Indeed, there …