Effective Utilization of Buffers in 802.11 Based WLAN

DEVANA RAJESWARI

Abstract

The main objective of this paper is to give high performance gain in order to achieve high throughput with less delay. In communication networks, buffers are used to accommodate short term packet bursts so as to less packet drops and to maintain high efficiency. Sizing buffers in wireless networks has fundamental issues arise when compared with wired, which leads to packet service times at different stations in WLANs being strongly coupled. the major presentation associated with the use of fixed size buffers in 802.11 WLANs and present dynamic buffer sizing algorithms that achieve significant performance gains and how the utility of proposed algorithm gives simulation results.

Figures (4)

Measured distribution of the MAC layer service time when there are 2 and 12 stations active. It can be seen that the mean service time changes by over an order of magnitude as the number of stations varies. Observe also from these measured distributions that there are significant fluctuations in the service time for a given fixed load. This is a direct consequence of the stochastic nature of the CSMA/CA contention mechanism used by the 802.11/802.11e MAC. the art which makes use of fixed size buffers, can easily lead to poor performance. . We recorded RTTs before and after one wireless station started to download a 37MByte file from a web-site. Before starting the download, we pinged the access point (AP) from a laptop 5 times, each time sending 100 ping packets. The RTTs reported by the ping program was between 2.6-3.2 ms. However, after starting This work is supported by Irish Research Council for Science, Engineering and Technology and Science Foundation Ireland Grant 07/IN.1/1901. the download and allowing it to continue for a while (to let the congestion control algorithm of TCP probe for the available bandwidth), the RTTs to the AP hugely increased to 2900-3400 ms. During the test, normal services such as web browsing experienced obvious pauses/lags on wireless stations using the network. Closer inspection revealed that the buffer occupancy at the AP exceeded 200 packets most of the time and reached 250 packets from time to time during the test. Note that the increase in measured RTT could be almost entirely attributed to the resulting queuing delay at the AP, and indicates that a more sophisticated approach to buffer sizing is required.

amongst TCP flows but can be readily resolved using 802.11le functionality by treating TCP ACKs as a separate traffic class which is assigned higher priority. With regard to throughput efficiency, the algorithms in this paper perform similarly when the DCF is used and when TCP ACKs are prioritized using the EDCA as in. Per flow behavior does, of course, differ due to the inherent unfairness in the DCF and we therefore mainly present results using the EDCA to avoid flow-level unfaimess. In this “paper is on TCP traffic since this continues to constitute the bulk of traffic in modem networks (80-90% of current Internet traffic and also of WLAN traffic), although we extend consideration to UDP traffic at various points during the discussion and also during our experimental tests.

Fig. shows example time histories of the buffer size and occupancy at the AP with a fixed buffer size of 400 packets and when the A* algorithm is used for dynamic buffer sizing. Note that in this example the 400 packet buffer never completely fills. Instead the buffer occupancy has a peak value of around 250 packets. This is due to non-congestive packet losses caused by channe! noise which prevent the TCP congestion window from growing to completely fill the buffer. Nevertheless, it can be seen that the buffer rarely empties and thus it is sufficient to provide an indication of the throughput when the wireless link is fully utilized.

where al and b1 are design parameters. Pseudo-code for this ALT algorithm is given in Algorithm. We can combine the eBDP and ALT algorithms by using the mean packet service time to calculate QeBDP as per the eBDP algorithm , and the idle/busy times to calculate gALT as per the ALT algorithm. We then select the buffer size as minfQeBDP ; gALT g to yield a hybrid algorithm. The basic design of the algorithm is to exploit the possibility of statistical multiplexing to reduce buffer sizes. Comparing with the results using fixed size buffers, we can see that the algorithm can achieve significantly smaller buffer sizes exists. It can be seen that in comparison with the use of a fixed buffer size the A* algorithm is able to achieve high throughput efficiency across a wide range of operating conditions while minimizing queuing delays. we further evaluate the algorithm when the wired RTTs are varied according to number of upload and download variation. Comparing these with the results of the eBDP algorithm we can see that the algorithm is capable of exploiting the statistical multiplexing where feasible. In particular, significantly lower delays are achieved with 10 download flows while maintaining comparable throughput efficiency.