Proposed an efficient multilevel dynamic bandwidth allocation (M-DBA) scheme for FiWi networks.

In recent years demand for high speed and robust wireless networks are increasing due to the growth of accessing social networks from mobiles and computer systems. The recent solution for this is used by Fiber-Wireless (FiWi) network which it can provide the high throughput for end users. FiWi supports the mobility of wireless networks as well as high bandwidth of optical networks. But the major problem with these networks is as the number of mobile user's increases, queuing delay is increasing at FiWi's Optical Network Unit (ONU). Increasing queuing delay resulted into decreasing the throughput performance of FiWi networks. Hence, it is required to have efficient queuing method at ONU to allocate bandwidth efficiently to get high speed performance of FiWi; one of such methods is called as Dynamic Bandwidth Allocation (DBA). To solve the problem of increasing delay in FiWi networks due to increasing number of users, connection patterns and traffic congestion, in this work we proposed a multilevel, multi-queue based DBA method, so it is called Multilevel Multi-queue DBA. The innovation of the proposed method compared to previous methods provides a complete multilevel framework for reducing delays to an acceptable level. In proposed method, every ONU is configured with three levels of priority queues. The real time data is placed into the queue which is having the highest priority queue, while the non-real-time data is placed into other two queues depending on the threshold values set for them. This algorithm works according to different TDMA slots (those are treated as levels) in order minimize the congestion in FiWi Networks, so it can help to minimize the queuing delay performance. The proposed method is simulated and evaluated using MATLAB simulator. From the simulation results, it is showing that proposed DBA method is having better performance in terms of queuing delay and packet delivery ratio as compared to existing methods. The performance of FiWi networks is increased by 60% for proposed DBA approach whereas throughput is increased by 55% as compared to existing DBA method. The simulation results also show that queuing delay is minimized by 30% as compared to existing DBA method. [ABSTRACT FROM AUTHOR]