Capacity analysis of densely deployed wireless LANs
Friday, 24 October 2014
Sal A, Electrum 1, Isafjordsgatan 26, Kista
Prof. Sunghyun Choi, Seoul National University, Korea
Prof. Olav Tirkkonen, Aalto University, Finland; Prof. Alexandre Proutiere, KTH Royal Institute of Technology, Sweden; Dr. Anders Furuskär, Ericsson Research, Sweden Main supervisor: Prof. Jens Zander, KTH Royal Institute of Technology
Doc. Ki Won Sung, KTH Royal Institute of Technology
Wireless LANs (WLANs) based on the IEEE 802.11 standard have become an integral part of today’s indoor wireless communication infrastructure. As WLAN deployments become more prevalent and densely deployed, the nodes in these WLANs start to create congestion and interference with each other. This congestion and interference fundamentally limits the performance of these coexisting WLANs. We analyze the capacity limits of such densely deployed WLANs.
We begin our analysis by investigating the suitability of the attributes of WLANs, namely their cooperative operation based on locally available information, for indoor high-capacity wireless access provisioning. We compare the cooperative class of wireless systems with another class of systems whose users behave selfishly.
Following this qualitative assessment, we perform a detailed, qualitative analysis of the capacity of densely deployed WLANs in terms of a number of key environmental and operational parameters. The indoor propagation environment has a significant influence on the congestion and interference that these coexisting WLANs exert on each other. Therefore we investigate the impact of propagation environment on the aggregate throughput of densely deployed WLANs. As WLANs are deployed in close proximity of each other, the transmissions in one WLAN start to influence the outcome of transmissions in other WLANs. The manner in which the access points are deployed, and the manner in which stations associate themselves with the available access points around themselves is shown to be an influential factor in the performance of these coexisting WLANs. Therefore, we investigate the impact of random versus planned access point deployment on performance of densely deployed WLANs. Similarly, we investigate the impact of stations associating with the access point with the strongest signal or with another sufficiently strong access point in their vicinity. Furthermore, we investigate the throughput of densely deployed WLANs when operating with bounded delay. More specifically we examine the case when the input traffic arriving at the transmitters are expected to reach their destination within a certain time period, thus the transmit queues cannot grow without bound and the system should operate at a stable point.
The indoor propagation environment, creates complex interference relationships between nodes in coexisting WLANs. These complex interference relationships are compounded by the node interactions dictated by the nonlinear algorithms in the IEEE 802.11 MAC protocol, thus the problem of estimating the performance of these coexisting WLANs by means of simple analytical models becomes difficult. In contrast, detailed packet level simulations provide accurate performance estimates, although such analyses are computationally expensive. Therefore we seek to provide a model to estimate the throughput of densely deployed WLANs based on empirical throughput results of detailed simulations of such densely deployed WLANs. In addition, in our effort to develop an empirical throughput model for densely deployed WLANs, we develop a measure which we call “cell congestion” to be able to order and compare different propagation environments, and an “effective density” concept which accounts for the influence of the propagation environment on the congestion and interference experienced by a WLAN deployment of a given density. We expect these concepts to be useful in improving the operation of WLANs to be able to meet the predicted increase in demand for capacity.