Performance Analysis And Improvement Of DCF In IEEE 802.11 Wireless Networks

With the improvement of information technology, IEEE 802.11 wireless network is developing dramatically. It has been widely used in our daily life. Wireless network makes people get rid of the shackle of cable, and make the access to the internet become easier no matter where and when. However, compared with the good performance of wired network, there is still a wide gap for wireless network. The reason for this is not only the inherent characteristics of wireless media, but the protocols in MAC layer. Thus, the performance analysis and improvement of MAC protocols are of great importance and practical value.This paper focus on DCF protocol in IEEE 802.11.And it enhances the Markov model in saturation condition which is proposed by Bianchi. Considering the number of retransmission times and the freezing state, I establish a new Markov model which is much closer to IEEE 802.11 standard and derive the formulas to calculate saturation throughput and delay in MAC layer. Then I analyze the saturation throughput performance and delay performance of the mainstream protocols such as IEEE 802.11a and IEEE 802.11b. Finally, I use simulation software NS2 to verify the accuracy of my result. In addition, I compare the performance in basic access mode and that in RTC/CTS access mode according to different number of active nodes and different size of data frame, and then I give the optimal thresholds to transfer to RTS/CTS mode based on the analysis.From the research on the Markov model in saturation condition, we get the relationship between collision probability p and the number of competing nodes n. We also track the number of nodes using the extended Kalman filter, take real time estimation on the number of nodes. From the simulation result, the estimation on the number of competing nodes is accurate.This paper also proposes the model for DCF protocol in non-saturation condition based on the analysis on saturation condition. Based on this model, this paper gives the throughput performances of IEEE 802.11a and IEEE802.11b in different saturation degree, and verifies the result by simulation using NS2.I also compare the performance in basic mode and that in RTS/CTS mode according to the number of active nodes and the size of the data, and give the thresholds to transfer to RTS/CTS mode.In IEEE 802.11 network, since AP and normal wireless nodes obey the same MAC protocol, there exists severe unfairness for uplink throughput and down link throughput, which result in AP becomes the bottleneck of WLAN. This paper propose a simple and effective priority method--Fair Bandwidth Allocation Scheme between Uplink and Downlink, FBASUD does not have to modify the existing terminals. It only needs to modify DCF protocol of AP a little in MAC layer, which can make AP has the relative high priority in the network. Thus, the fair bandwidth allocation can be achieved. From the simulation result, this mechanism can improve the unfair situation obviously and solve the bottleneck problem in WLAN.In order to solve the time-possession unfairness problem for the nodes with different data rate, this paper proposes a time fairness algorithm based on Markov model-- Fair Bandwidth Allocation Scheme between Highspeed and Lowspeed, FBASHL algorithm takes the time that the node possesses the channel when it transmits data, not the probability that the node can access the channel, as the measurement of the fairness. Every node maintains the fairness of accessing channel in WLAN through adjusting the number of times to compete for wireless channel. Using NS2 to simulate, we can verify that this protocol can effectively solve the fairness problem for the nodes with different data rate, thus, improve the throughput performance.Based on the Markov model that this paper improved, we derive the VoWLAN capacity in basic access mode and that in RTS/CTS mode in 802.11b network. And I use NS2 to simulate the network to verify the effectiveness of this model. Based on this analysis, I discuss the methods to improve the VoWLAN capacity by compressing the header, enlarging the size of the voice frame, improving the priority for AP and so on.