Research On Access Algorithms And QoS Performance In Wireless Network

Wireless network as an important part of personal computer communication has been expanding a new wave in the field of mobile computing. It would be applied widely into our social life in the future. As a research hotspot at present, access algorithm in wireless network and its quantitative performance analysis are regarded as significant theoretical fundamentals and supporting technology of studies and application in computer network and communication system.The development of wireless network is supposed to be an innovation process. With the advancement of broadband wireless access network and the third generation mobile communication, theoretical basis of access protocol and correlative algorithm in wireless network should be studied and improved further. However, the investigation to quality of service (QoS) in wireless access network just lies in start stage. Meanwhile a feasible and reliable solution has not been proposed in standard protocol. Additionally, the quantitative analysis of QoS is a new challenge in wireless access network.The paper mainly conducts quantitative analysis of and systemic research on Backoff algorithm, channel allocation algorithm, AP load balancing algorithm, access control algorithm and GPRS access algorithm in wireless access network.The main works of the paper include the following five parts:In the first part, a novel p' -Persistent Backoff Algorithm (P_PBA) based on IEEE 802.11 DCF is proposed and quantitative analysis has been carried out. The P_PBA algorithm can adjusts contention window size dynamically according to network load in order to decrease collision probability effectively and perfects CSMA/CA protocol in IEEE 802.11. The simulation results indicate P_PBA, while decreasing the frame delay, can offer a higher throughput, thus enhancing network performance significantly.In the second part, we evaluate the performance of occupied channel and employ "borrow" allocation technology based on channel unreliability and user mobility. Further channel allocation algorithm supporting real-time traffic and computation measure for queue buffer is proposed respectively, and a corresponding model is put forward for QoS quantitative analysis. The simulation results demonstrate that the improved channel allocation strategy performs more effectively performance and stabilization than the standard access algorithm in IEEE 802.11.In the third part, we propose an efficient load balancing method concerning both signal strength and each AP's current load on the basis neighbor graph. And each AP interacts with distributed system to decrease access delay. The simulation results indicate that the whole access performance in wireless network proves to be more efficient than it in traditional method when taking account for signal strength and AP current load synthetically.In the fourth part, a systemic model is created based on of G/M/c/∞queue system, which is regarded as a theoretical basis of optimal strategy. In the discussion of the benefit of service provider and the traffic priority of access user in wireless network, a novel strategy control model and access control algorithm OPMAC based on Markov process are proposed as a supplement to IEEE 802.11e standard protocol combining with virtual AP technology. Then correctness of the model is validated by quantitative computing and theoretical analysis. Finally, the comparison of some simulation experiments proved that the new algorithm can effectively provide different QoS to traffics with diverse requirements.The fifth part studies call admission in GPRS and channel modeling, and further puts forward an analytical model for throughput Computation and bandwidth allocation and a novel GPRS access algorithm based on ON/OFF state in mobile wireless network. Moreover, we accomplish QoS quantitative analysis and computation of Markov process. Eventually, some analysis formulas are obtained, which are of significant value both theoretically and practically. Computations and experiments indicate that the numerical results from analysis formula are consistent with our simulations.