A control theoretic approach to achieve proportional fairness in 802.11e EDCA WLANs
Introduction
Enhanced Distributed Channel Access (EDCA) was proposed in IEEE 802.11e-2005 standard to support QoS enhancement and service differentiation for WLAN applications [1]. It extends the basic Distributed Coordination Function (DCF) by classifying traffic flows into four different Access Categories (ACs), namely voice, video, best-effort and background. Traffic with higher QoS requirements, e.g. shorter delay deadline, is assigned a higher priority, and hence, on average, waits for less time before being sent to the channel. This mechanism is beneficial for high-priority traffic. Compared to the DCF, EDCA sacrifices the performance of low-priority traffic to some extent to provide QoS support for high-priority traffic. When the network is saturated with a large proportion of high-priority flows, an extremely unfair scenario will appear, in which the channel will be almost completely occupied by high-priority flows, e.g. VoIP or video streaming flows, however low-priority traffic, such as email or web browsing data, will suffer severe starvation.
Resource allocation in EDCA WLANs has therefore been the subject of considerable interest. The objective is to seek for a fair allocation of network resources (e.g. throughput, airtime and etc.) amongst different traffic types, and meanwhile, guarantee the specific QoS requirements and service differentiation. This paper considers proportional fair allocation of station throughputs amongst ACs for provision of distinct average delay deadlines and priority parameters. The 802.11e EDCA standard specifies four contention parameters to distinguish priority levels, which are minimum Contention Window (CWmin), maximum Contention Window (CWmax), Arbitration Inter Frame Space (AIFS) and maximum Transmission Opportunity (TXOP). A set of default values for the four parameters are recommended in the standard for each physical (PHY) layer supported by 802.11e. As the default values do not take into account the varying WLAN conditions, and thus lead to suboptimal performance and no fairness guarantees, in this paper we find the optimal CWmin value that leads to proportional fair allocation of station throughputs while assuming AIFS and TXOP taking the recommended values and . The optimal CWmin value corresponds to an optimal station attempt probability which is derived from the proportional fairness analysis.
In order to implement the derived proportional fair allocation in practice, a centralised adaptive approach which uses multivariable state-space control theory is then proposed. The WLAN is represented as a discrete multi-input multi-output (MIMO) linear time-invariant (LTI) state-space model. A state feedback control method, the Linear Quadratic Integral (LQI) control, is used to tune the CWmin value to drive the station attempt probability to the optimum so as to maintain a fair throughput allocation. We have demonstrated in simulations that the proposed control approach is adaptive to general network scenarios with high accuracy and fast convergence speed. To our knowledge this might be the first time that a closed-loop control system is designed for EDCA WLANs to achieve proportional fairness amongst ACs.
The remainder of this paper is organised as follows. Section 2 gives a comprehensive review on the state-of-the-art research on fairness and control theory approaches to solve network problems. Section 3 presents the theoretical analysis to derive the optimal station attempt probability which leads to proportional fair allocation of station throughputs given the constraints on the average delay deadlines for different ACs. Section 4 describes a centralised adaptive control approach which can realise the proportional fair allocation derived from Section 3 in real networks. Section 5 evaluates the performances of the fairness algorithm and the proposed centralised control approach, and Section 6 concludes the paper.
Section snippets
Fairness
Fairness has been the subject of a considerable body of literature on 802.11 WLANs [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19]. The unfairness behaviours may be caused by a number of factors, e.g. hidden terminals, exposed terminals, capture, uplink/downlink unfairness, asymmetric radio conditions and multiple data rates and etc., which have been investigated in [2], [4], [5], [6], [7], [8], [11], [13]. There also exist distinct fairness
Network model
We consider a single-hop 802.11e EDCA WLAN with one AP and n client stations, as depicted in Fig. 1. The channel is assumed to be error-free for all supported PHY rates. Traffic flows are classified into N different ACs. We assume that each client carries flows of a single AC, so there are no virtual collisions in our setup. The number of stations in the ith AC is ni. The total number of stations is thus . The analysis can be readily generalised to encompass situations where client
Centralised closed-loop control approach
In this section, we design a centralised adaptive control approach to implement the desirable proportional fairness in real networks. Based upon the analysis in Section 3, the proportional fairness is achieved when the station attempt probability parameter α reaches its optimum value α*. The variable α is only determined by the minimum contention window Wi with AIFS and TXOP taking the recommended values and . Our approach uses a multivariable closed-loop control system to tune W to
Throughput and delay performance
The main objective of this work is to achieve proportional fair allocation of station throughputs while satisfying specific delay constraints of different ACs. To verify if the proposed fairness algorithm meets this objective, we first evaluate the throughput allocation and delay performance. The results are obtained using Matlab based on the throughput and delay analysis in Sections 3.2, 3.3 and the proportional fairness algorithm described in Section 3.4. As the throughput and delay analysis
Conclusions
This paper considers using a closed-loop control approach to achieve proportional fair allocation of station throughputs in a multi-priority EDCA WLAN. The optimal station attempt probability that leads to proportional fairness is derived given the average delay deadline constraints of different ACs present in an WLAN. To achieve the desirable proportional fairness, a centralised adaptive control approach is proposed. The WLAN is represented as a discrete MIMO LTI state-space model. The LQI
Acknowledgements
This work was conducted as part of the project ISS-EWATUS (issewatus.eu) and has been financially funded by European Commission through the FP7 program (grant agreement no: 619228).
References (38)
- et al.
A simple but accurate throughput model for IEEE 802.11 EDCA in saturation and non-saturation conditions
Comput. Netw.
(2011) - et al.
Bio-inspired self-adaptive rate control for multi-priority data transmission over WLANs
Comput. Commun.
(2014) - et al.
Performance analysis under finite load and improvements for multirate 802.11
Elsevier Comput. Commun. J.
(2005) - IEEE 802.11Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) specifications: Amendment 8: Medium Access...
- et al.
MACAW: a media access protocol for wireless LANs
Proceedings of the ACM SIGCOMM
(1994) - et al.
A practical cross-layer mechanism for fairness in 802.11 Networks
Mob. Netw. Appl.
(2005) - et al.
Improving throughput and fairness by reducing exposed and hidden nodes in 802.11 networks
IEEE Trans. Mob. Comput.
(2008) - et al.
Sniffing out the correct physical layer capture model in 802.11b
Proceedings of the IEEE ICNP
(2004) - et al.
TCP fairness in 802.11e WLANs
IEEE Commun. Lett.
(Nov. 2005) - et al.
MAC layer channel quality measurement in 802.11
IEEE Commun. Lett.
(2007)
Achieving per-flow and weighted fairness for uplink and downlink in IEEE 802.11 WLANs
EURASIP J Wirel. Commun. Netw.
AP association for proportional fairness in multirate WLANs
IEEE/ACM Trans. Netw.
Charging and rate control for elastic traffic
Eur. Trans. Telecommun.
Proportional fair throughput allocation in multirate IEEE 802.11e wireless LANs
Wirel. Netw.
Proportional fairness for QoS enhancement in IEEE 802.11e WLANs
Proceedings of the IEEE Conference on Local Computer Networks
Optimal CWmin selection for achieving proportional fairness in multi-rate 802.11e WLANs: test-bed implementation and evaluation
Proceedings of the First International Workshop on Wireless Network Testbeds, Experimental Evaluation and Characterization
Optimal configuration of 802.11e EDCA for real-time and data traffic
IEEE Trans. Veh. Technol.
Dynamic contention window control scheme in IEEE 802.11e wireless LANs
Proceedings of the IEEE VTC
A new scheme to achieve weighted fairness for WLAN supporting multimedia services
IEEE Trans. Wirel. Commun.
Cited by (6)
Formal Modeling and Verification of EDCA Based on Probabilistic Model Checking
2020, Proceedings - 2020 IEEE 22nd International Conference on High Performance Computing and Communications, IEEE 18th International Conference on Smart City and IEEE 6th International Conference on Data Science and Systems, HPCC-SmartCity-DSS 2020Quality of Service Enhancement in Wireless LAN: A Systematic Literature Review
2019, MACS 2019 - 13th International Conference on Mathematics, Actuarial Science, Computer Science and Statistics, ProceedingsImproving QoS in Wireless Mesh Networks Using Heterogeneous and Hybrid Scheduling Design Approaches
2019, 2nd International Conference on Computer Applications and Information Security, ICCAIS 2019Adaptive access mechanism with delta estimation algorithm of traffic loads for supporting weighted priority in IEEE 802.11e WLANs
2019, Journal of Ambient Intelligence and Humanized ComputingAccess mechanism supporting weighted throughput fairness of multiple priority levels
2018, IPPTA: Quarterly Journal of Indian Pulp and Paper Technical AssociationThroughput Optimization of multiple association-based EDCA software defined dense WLAN
2018, Gaojishu Tongxin/Chinese High Technology Letters