Efficient Bandwidth Sharing Using Adaptive Token Bank Fair Queuing Algorithm in Wireless Networks Using a Cross Layer Approach

摘要

Wireless networks are characterized by time-varying nature as well as scarce resources. The data rate of wireless radio frequency channel is limited by Shannon's capacity law. Traditional methods including opportunistic scheduling algorithms segregates packet scheduling and resource allocation is inefficient. Though fairness and throughput are inversely related, a wise compromise between the two will result in acceptable QoS (Quality of Service) levels with comparable throughput. To improve the utilization of limited capacity, packet scheduling as well as resource allocation should be inter linked for maintaining throughput as well as fairness issue. This type of interaction between physical and medium access layer refers to cross-layer resource scheduling. The study presents a scheduler for packet scheduling and resource allocation taking queue and channel states in to account. In first level of scheduling, users to be given chance are selected based on certain conditions which include backlogged queues, fairness and delay in users. The second level of scheduling includes resource allocation for selected users using OFDM (Orthogonal Frequency Division Multiplexing) technique. Token Bank Fair Queuing (TBFQ) is based on leaky-bucket mechanism with each flow keeps track of number of tokens borrowed from or given to the token bank using a counter. Tokens overflowing the pool are stored in the bank. Priority of a connection in borrowing tokens is based on the counter value and token generation rate. As the flow is serviced, the value of counter varies which changes the priority level for further processing so that starving flows are also given service in this algorithm. The flow which has highest priority is called for amount of budget needed and based on the budget, sub carrier is allocated using OFDM. The channel conditions are measured using SNR (Signal to Noise Ratio) and given as feedback so that modulation scheme is varied as per the channel conditions for effective performance. This Adaptive TBFQ (ATBFQ) scheduling scheme can improve the wide spreading of Wi-max technology for future enhancements.