Supporting QoS in IEEE 802.11e wireless LANs over fading channel

Abstract

Transferring real-time traffic such as voice and video over wireless LAN networks (WLAN) requires stringent delay and jitter requirements. Recently IEEE 802.11e standard has been emerged to support QoS in WLAN. One of the methods to provide QoS in this standard is Enhanced Distributed Channel Access (EDCA) which benefits form the concept of traffic categories. However, EDCA is a contention based method; therefore it can not guarantee strict QoS required by real-time services without proper network control mechanisms. In this paper, we analyze the effect of loss and delay caused by fading channel on EDCA performance. Then, we propose a modification to the media access scheme, called CAFD (Collision Avoidance with Fading Detection) to elevate the performance against channel failures. Moreover an adjustment for the maximum number of retransmissions is proposed to maintain the delay and jitter requirements of the real-time traffic. The performances of the proposed methods are evaluated by simulations.

Introduction

One of the channel access mechanisms in IEEE 802.11 is Distributed Coordination Function (DCF). DCF is based on CSMA/CA.¹ In this mechanism, a station waits for a quiet period in wireless media, and then begins to transmit data while detecting collisions. The time lapse between successive carrier senses, when channel is occupied, is given by a back-off counter which has an initial random value within a predetermined range. DCF does not support any type of priority access to the wireless medium. As a consequence, DCF provides only best-effort service, and there is no mechanism to provide better service for real-time traffic. Another drawback against QoS provisioning is that the 802.11 MAC does not specify any admission control mechanism. This implies that under heavy traffic load the performance deteriorates in an uncontrollable manner. To address these deficiencies, the “IEEE 802.11 Task Group E” recently has proposed a new contention-based channel access method called EDCA for the IEEE 802.11e standard [1].

Many research efforts have been done to study the IEEE 802.11 DCF performance, by both of analysis and simulation. Most of them assume the ideal channel condition, which means that the packet corruptions are only due to collisions [2], [3], [4].

In [5], based on the IEEE 802.11 DCF, a novel scheme named DCF_f is proposed to improve the performance of Wireless Local Area Network (WLAN) in fading channel.

Several priority studies have been reported in the literature for EDCA. In [6] three approaches are proposed; first, dynamic ReAllocative Priority (ReAP) scheme, wherein the priorities of packets in the MAC queues are not fixed, but keep changing dynamically. It uses the laxity and the hop length information to decide the priority of the packet. Second, Adaptive-TXOP² (A-TXOP), that involves modifying TXOP interval dynamically based on the packets in the queue. Third, TXOP-sharing, that uses TXOP to transmit to multiple receivers in order to utilize the TXOP interval completely.

The method mentioned in [7] is based on the network condition. It gradually lessens the contention window, instead of resetting the initial window value to ${\mathit{CW}}_{\mathit{min}}$ directly, to avoid continuous collisions.

The authors of [8] have proposed a method called MEDCF.³ This method adjusts the contention window dynamically according to the average collision rate, which is an indication of the traffic load.

The proposed model in [9] distinguishes internal collision from external collision. It also differentiates the two cases when the backoff counter decreases, i.e. an arbitration interframe space (AIFS) period after a busy duration and a time slot after the AIFS period.

The authors of [10] have proposed a three-dimensional Markov chain model for the EDCA mode, and based upon this Markov model, they have computed the throughput that different traffic classes can sustain, and the distribution of the channel-access delay that each head-of-line (HOL) packet experiences when the network is heavily loaded.

In [11] a method has been proposed to modify the contention behaviors of voice and data nodes. For a contending voice (or data) node, after waiting for the channel to be idle for an AIFS[AC voice] (or AIFS[AC data]), the node will send a black burst to jam the channel. The length of the black burst is equal to the node backoff timer.

In [12], an analytical model is proposed to evaluate the performance of the 802.11e in the saturated case.

The authors in [13] analyze the timing behavior of the EDCA function, when it is used to support real-time traffic.

In [14], the authors evaluate, through simulations, the benefits derived from the adoption of the IEEE 802.11e standard in delivering multimedia over WLANs.

To contemplate call admission control in EDCA, Ref. [15] proposes two mechanisms. The idea is, when accepting a new real-time flow, the admission control algorithm considers its effect on the channel utilization and the delay experienced by existing real-time flows, ensuring that the channel is not overloaded and the delay requirements are not violated. At the same time, the rate control algorithm allows the best effort traffic to fully use the residual bandwidth left by the real-time traffic. In this study the channel is assumed to be perfect with no packet loss due to channel fading.

In this paper, first we develop an analytical model to derive an average delay and loss estimate for 802.11e over fading channel. This study leads us to propose a modification to the MAC protocol to bring about better performance. As the second contribution, an adjustment for the maximum number of retransmissions is proposed to maintain the QoS requirements of the real-time traffic sources.

The remainder of this paper is organized as follows. In Section 2, we give a brief introduction to IEEE 802.11e EDCA. The effect of channel failures on EDCA performance is explained in Section 3. The average transmission delay in a fading channel is analyzed in Section 4, and the loss probability is evaluated in Section 5. We then present our proposals for MAC protocol and maximum number or retransmissions adjustment in Section 6. In Section 7, the performances of the proposed schemes are evaluated through simulation studies. Finally, Section 8 concludes this paper.