Understanding Fairness and its Impact on Quality of Service in IEEE 802.11

Understanding Fairness and its Impact on Quality of Service in IEEE 802.11 Michael Bredel, Markus Fidler Multimedia Communications Lab, TU Darmstadt Abstract— The Distributed Coordination Function (DCF) aims at fair and efficient medium access in IEEE 802.11. In face of its success, it is remarkable that there is little consensus on the actual degree of fairness achieved, particularly bearing its impact on quality of service in mind. In this paper we provide an accurate model for the fairness of the DCF. Given M greedy stations we assume fairness if a tagged station contributes a share of 1/M to the overall number of packets transmitted. We derive the probability distribution of fairness deviations and support our analytical results by an extensive set of measurements. We find a closed-form expression for the improvement of long-term over short-term fairness. Regarding the random countdown values we quantify the significance of their distribution whereas we discover that fairness is largely insensitive to the distribution parameters. Based on our findings we view the DCF as emulating an ideal fair queuing system to quantify the deviations from a fair rate allocation. We deduce a stochastic service curve model for the DCF to predict packet delays in IEEE 802.11. We show how a station can estimate its fair bandwidth share from passive measurements of its traffic arrivals and departures. I. INTRODUCTION The Distributed Coordination Function (DCF) specifies a randomized access procedure for the shared wireless medium in IEEE 802.11. The target is to divide resources fairly among an unknown number of stations while minimizing access delays and maximizing overall throughput. Originating from the basic ALOHA access scheme significant progress has been made regarding throughput and stability of today’s Medium Access Control (MAC) protocols [1], [2]. The issue of per-flow fairness is however, still under debate and different studies do not agree in their conclusions, e.g. [3], [4]. An important aspect of fair scheduling is the attainable quality of service. While long-term fairness ensures a certain average throughput the issue of short-term fairness has tremendous impact on individual packet delays [3], [5], [6]. Centralized fair scheduling algorithms on the other hand are very well understood today. The pioneering Generalized Processor Sharing (GPS) model [7] assumes a weighted re- source allocation that is perfectly fair on any time scale. To date, a variety of packet-by-packet implementations exist that emulate GPS closely, such as Weighted Fair Queuing. Distributed emulations are proposed in [8], [9], [10] with the aim to implement fair scheduling in the DCF. Models for fair packet scheduling are derived e.g. in [11], [12], [13], [14]. These models define error terms that specify the worst-case deviation of a packet scheduler from an ideal GPS system. Moreover, the GPS model and the calculus for network delay [15] gave rise to the important concept of deterministic service curve [7] that is the foundation of today’s network calculus [16], [17]. Recently, significant progress has been made towards the formulation of stochastic service curves, see [18], [19], [20] and references therein. These models are used in [21], [20], [22] to derive service curve representations of wireless links with a focus on channel outages that are due to fading and interference. Modeling random medium access is, however, an open challenge. In this paper we analyze the fairness of the DCF and the impacts on quality of service at a single radio channel. Our contributions are as follows. First, we derive closed- form solutions for the conditional distribution P[K = k|l] that a contending station transmits k packets given a tagged station transmits l packets within the same time interval. This characterization of fairness turns out to be comprehensive and versatile, e.g. the well-known fairness index by Jain [23] follows readily. We substantiate our analytical findings using an extensive baseline set measurements that we conducted in a shielded and unechoic room and of OmNet++ simulations. Second, we view the DCF as emulating the GPS policy. We formulate a recursive model for packet departure times coined DCF clock that is subject to well-defined random error terms. Based on the distribution of packet inter-transmissions we derive a stochastic service curve model for the DCF. Finally, we show how a station can obtain reliable estimates of the fair rate from passive measurements of its arrivals and departures. The remainder of this paper is structured as follows. In Sect. II we discuss related work on fairness in IEEE 802.11. In Sect. III we elaborate on our controlled evaluation environment. We compare our set of baseline measurements to OmNet++ simu- lations as well as to related studies. In Sect. IV we develop a model of the DCF and derive closed-form expressions for the fairness. In Sect. V we view the DCF as emulating GPS and derive the DCF clock

…(Full text truncated)…

This content is AI-processed based on ArXiv data.