For Full-Text PDF, please login, if you are a member of IEICE,|
or go to Pay Per View on menu list, if you are a nonmember of IEICE.
Mitigating Throughput Starvation in Dense WLANs through Potential Game-Based Channel Selection
Bo YIN Shotaro KAMIYA Koji YAMAMOTO Takayuki NISHIO Masahiro MORIKURA Hirantha ABEYSEKERA
IEICE TRANSACTIONS on Fundamentals of Electronics, Communications and Computer Sciences
Publication Date: 2017/11/01
Online ISSN: 1745-1337
Type of Manuscript: Special Section PAPER (Special Section on Smart Multimedia & Communication Systems)
Category: Communication Systems
densely deployed WLANs, flow-in-the-middle starvation, channel selection, potential game,
Full Text: PDF(2.2MB)>>
Distributed channel selection schemes are proposed in this paper to mitigate the flow-in-the-middle (FIM) starvation in dense wireless local area networks (WLANs). The FIM starvation occurs when the middle transmitter is within the carrier sense range of two exterior transmitters, while the two exterior transmitters are not within the carrier sense range of each other. Since an exterior transmitter sends a frame regardless of the other, the middle transmitter has a high probability of detecting the channel being occupied. Under heavy traffic conditions, the middle transmitter suffers from extremely low transmission opportunities, i.e., throughput starvation. The basic idea of the proposed schemes is to let each access point (AP) select the channel which has less three-node-chain topologies within its two-hop neighborhood. The proposed schemes are formulated in strategic form games. Payoff functions are designed so that they are proved to be potential games. Therefore, the convergence is guaranteed when the proposed schemes are conducted in a distributed manner by using unilateral improvement dynamics. Moreover, we conduct evaluations through graph-based simulations and the ns-3 simulator. Simulations confirm that the FIM starvation has been mitigated since the number of three-node-chain topologies has been significantly reduced. The 5th percentile throughput has been improved.