Autonomic resource management in IEEE 802.11 open access networks

Sammanfattning: This work is inspired by the autonomic networking paradigm. It investigates the issue of making network resource management mechanisms in the access part of wireless networks self-aware and self-configuring. Open access networks (OAN) have private WLAN cells made available for public use and enable bypassing mobile users to profit from continuous coverage, allowing private and public subscribers to share a common infrastructure. Resources are network parameters, e.g. bandwidth, capacity, that have to be fairly and accurately allocated, distributed, and managed in order to optimize performance and ensure proper and fair operation. Analyzing and exploring existing multimedia session and resource management mechanisms on the wireless part is a crucial step for determining their limitations and shortcomings. A new way of making OANs self-aware is proposed and a self-configuring mechanism which has a competitive advantage in the qualitative sense over already existing mechanisms is designed and developed. The amount of network resources as well as the timing and form of their availability are decisive for enabling service functionality and providing a sufficient quality of service level. Our area of discourse is the "last- mile" part of an 802.11 network; i.e. the part between the access point and the mobile node. In OANs, maintaining and updating rates for different flows used by a single terminal traversing many adjacent APs is a challenge in itself. It is solved by using several key design factors. Monitoring the environment locally and updating the designed information structures leads to network self-awareness. Next, a high-level mechanism that aggregates information from several protocols including CARD (Candidate Access Router Discovery), MIP (Mobile IP), and SIP (Session Initiation Protocol) is defined. It specifies the interactions between the mobile terminal and the access point controller. Then capacity maintenance and rate allocation issues at the two end-points of the last-mile are analyzed. The overhead in OANs required for the launching or reconfiguration of a real-time session can be significantly reduced by a proposed algorithm which performs parameter injection into a multimedia data stream instead of following the classical negotiation path. A solution to change the session characteristics of an application that adapts to varying network conditions was developed, including a QoS solution with SIP and MIP to optimize the session characteristics by taking into account the changed conditions in the network. With this approach, reduced session signaling delay and low jitter values are achieved as implementation results show. The thesis is concluded by considering the general issues: self-aware node architectures, self-configuring resource management.