Research On Efficient And Fair Access Control Technologies In Wireless Mesh Network

With the development of wireless communication new services, the performance requirements for wireless local area network (WLAN), wireless mesh network (WMN) and cellular wireless network are expected higher demand, such as transmission rate, spectrum efficiency and quality of service (QoS). Due to the features of wireless, mobile and multi-hop, it is difficult for traditional protocol mechanisms designed for single-hop network to meet the performance requirements while desiging WMN. As the standard of WMN, one distinct feature of IEEE802.11s is the fact that the mesh network is implemented at link layer, relying on MAC address rather than IP address for its mechanisms. IEEE802.11e EDCA is still as the baseline of MAC mechanism in IEEE802.11s standard draft. However, facing the high QoS demand of real-time streaming media traffic, it encounters the great challenges on performance, including poor fairness, without traffic differentiation, low utiltiy of bandiwidth and so on. This thesis aims at the different QoS demand of different types of traffic, such as voice, streaming media traffic and data traffic, addresses the related issues of MAC enhanced technologies in WMN, including access control, congestion control and frame aggregation by using pricing strategy, game theory and queue theory. The goal is to to improve the MAC efficiency and obtain the good balance among the channel resource utility, QoS and fairness.Firstly, taking into account the characteristics and limitation condtions of WMN, we investigated the congestion control scheme, and mainly focused on finding the solutions to two key problems, The one is to design the efficient congestion control scheme to alleviate local congestion condition, and the other is to find the good balance point between QoS guarantee and fairness among different types of traffic. Based on pricing mechanism, we proposed two hop-by-hop congestion control schemes for single-channel WMN. The first scheme is based on the congestion control framework proposed by IEEE802.11s standard draft, and adopts dynamic pricing function and new rate control model in process of local congestion monitoring and local rate control, in order to improve the efficiency of congestion control and network performance, for example, to increase the system throughput. The MAC time constraint and binary interference model constraint are considered in the second congestion control scheme. The goal is to obtain proportional fairness for the sum of each flow utility. The transmission failure ratio is introduced into congestion price as one of the criteira of congestion control. The rate control is implemented via changing the modulation mode of physical layer. Additionally, in order to guarantee the QoS requirement and fairness of each flow, the intra-node flow scheduling scheme is taken into account. Different from one-hop flow control scheme in traditional link layer and end-to-end congestion control scheme in traditional transport layer, the two congestion control schemes proposed by this thesis deal with rate control of each hop node along the flow path. It is a generalized congestion control problem. Simulation and analyse results show that our two schemes are efficient, which can alleviate congestion and starvation state of traffic with low priority effectively, while improve some performance, such as the average delay and average throughput.Secondly, we address the access control issue in MAC and analyze the unfairness problem in current IEEE802.11e MAC scheme, that is, the higher priority traffic flows intend to severely deprive the access opportunity of lower priority traffic flows due to the capture effect. Adopting traffic type differentiable and dynamic priority methods, we focus on how to balance the conflict between QoS demand of diversified traffic and fairness, and prevent higher priory traffic extremely depriving channel resource. A single criterion fair access control algorithm based on bandwidth occupied time proportion and a multi-criteria fair access control lgorithm with dynamic priority are proposed by this paper. By overcoming the limitation of fixed priority defined in802.11e EDCA, we dynamically tune the priority of different type of traffic to achieve a fair access of the bandwidth resource among different traffic flows. The simulation results indicate that the two algorithms proposed by this thesis can efficiently improve the unfair situation of lower priorty accessing the channel and meet the QoS requirement of different types of traffic, thereafter improve other network performance, such as throughput and the ratio of packet loss.Frame aggregation technology has recently been applied to various fields, for example, Ethernet, optical fiber nework, the third generation network and WMN. Frame aggregation mechanism of MAC layer is adopted by IEEE802.1In standard which aimed at improving transmission rate. It is regarded as one of the most efficient method to improve the throughput of MAC and bandwidth utility with the least modification for current protocol. Since most current research results of frame aggregation mechanism seldomly consider the traffic differentiation issue, we focus on the study of frame aggregation technology from the aspect of differentiating the Qos requirement of different types of traffic and improving the performance of system. Two frame aggregation schemes based on queue model are proposed. The first scheme adopts weight fairness and two-level buffer schedule ideas. The aggregation weight factor of each queue is dynamically adjusted based on the packet overtime urgent factor defined to satisfy the delay requirement of each QoS type of service. A mathematical analysis model under traffic differentiation scene is set up and a centralized parameter substituting and approximation method is applied to solve the problems. Additionally, the relationship of main performance parameters, including mean length of queue and mean waiting time, and system load is quantificational analyzed. The second scheme is a frame aggregation scheme based on static pricing and utilty. By using the average delay as main QoS criterion, we investigate the optimizationthe problem of price and network utility from the aspects of user utility and network utility. The relationship between the packet arrival rate and the average delay is analyzed. Numerical anaysis and simulation results show that our proposed frame aggregation schemes can improve the system average throughput and decrease the end-to-end delay.Finally, we address the problem of access admission control game between STA and mesh AP (MAP) in WMN. The access admission control mechanism in WMN and WLAN plays an important role in achieving fair resource schedule and load balance. Both STA and MAP will encounter the problem of how to select a proper access point during the process of accessing a network. The goal of both players is to seek the maximam of their utility. The optimal resource allocation can be attained via object restriction and decision function. It can be modeled as a game process of optimal admission control strategy decision, and the decision on whether an access request can be admitted or not was determined by decision factors. Based on pricing mechanism and game theory, we study from both the users'and the service providers'point of view, and formulate the access admission control process between STA and MAP as a nonzero-sum, noncooperative and mixedstrategy game. Three decision factors and mixed strategy probability function for MAP and STA were defined. The calculate steps of the weight value for each factor using Analytic Hierarchy Process (AHP) were discussed.