Research On Cross-layer Optimization Of The Wireless Mobile Networks Based On QoS Guarantee

Now days, In wireless mobile networks QoS of services in fairness, real-time is hard to get the effective guarantee, limit the development of wireless mobile networks, the reasons analyzed in QoS guarantee plane as follow: (1) the existing QoS mapping methods based on static mapping are unable to effectively satisfy users'different kinds of QoS requirements including real-time; (2) the actual performance of traditional scheduling algorithms would be seriously weakened in wireless channel environment, thereby unable to meet service requirements of real-time and fairness; (3) the traditional TCP congestion control is hard to judge the real reason causing network congestion accurately, often leads to unnecessary congestion avoid and window narrowed, increase network delay and retransmission, impair the consistency and fairness of wireless services. The cross-layer design method as a key technology of the next generation wireless mobile communication system, can take advantage of the cross-layer optimization to improve QoS guarantee performance, and provide new method for the research of this article.To resolve the mobility and time-vary of wireless mobile networks, and guarantee the fairness and real-time of service QoS, the article proposed three cross-layer optimization methods from aspects of QoS mapping, packet scheduling, and congestion control. The main contents included:1. To propose GRQM cross-layer mapping algorithm based on the SLA mechanism. Aiming at the shortages of layer-by-layer QoS mapping method, the cross-layer gray correlation analysis mapping (GRQM) algorithm sought the closest degree degrees between users' experience QoE and network QoS bearing index, to realize the accurate QoS mapping. Through theoretical and simulation analysis: this algorithm can improve the limitations of the static or simple linearity QoS mapping and layer-by-layer QoS mapping, made QoS mapping parameter reflects the user's QoS requirements more accurately.2. To design a cross-layer allocation algorithm of time slot for real-time services. This algorithm took account of the real-time requirements, channel state and queue length, and according to delay and real-time throughput parameters estimated the packet's scheduling priority, and allocated time slot according to it. Through theoretical and simulation analysis: on the basis of ensuring business throughput, this algorithm can also effectively prevent the users with higher priority and better channel condition to occupancy too much resources, ensure services of users with lower priority and worse channel conditions, thereby improve fairness and throughout of system.3. To put forward a cross-layer congestion control scheme based on congestion grade (GradeCC). This scheme established cross-layer Snoop agent in receiving end, and judged network congestion degree based on the data transmission parameters from link layer, it would submit the explicit messages to the sending end, so it can effectively distinguish the network packet loss caused by congestion or error code, and took accurate window control to resolve network congestion through. Through theoretical and simulation analysis: the congestion grading can make the sending end quickly and accurately adjust the window size accord with congestion grade of the network; does not reduce packet transmission rate as far as possible when error code causing packet loss, thereby guarantees the network throughput and the fairness and real-time of services.