6. Conclusions
Interference is a fundamental issue in wireless networks and it affects the aggregated throughput of a network. In this paper we have characterized the power interference in IEEE 802.11 CSMA/CA based networks using DA. An improved Attacking Case metric that quantizes the severity of interference has been proposed using the Link-Interference Graph, Transmitter-side Protocol Collision Prevention Graph, and Receiver-side Protocol Collision Prevention Graph. This metric differs from Liew’ Attacking Case metric proposed in [1] as the original metric only addresses networks using OAs. Our improved Attacking Case metric is meant for networks using DA but it can also be used in networks using OA. Our proposed Attacking Case metric is pertinent as there are no metric available currently to quantize the severity of interference in IEEE 802.11 CSMA/CA based networks using DA. It was also found that interference is tied with Attacking Case, thus reducing Attacking Case can result in an increase of throughput. The reduction of Attacking Case can be achieved by the usage of strategies such as DA, transmission power reduction, or DA with transmission power reduction. The relationship between Attacking Case and the throughput of a network is worth to be studied; if there is a statistically strong relationship between these two, a model could be built which is useful to predict the throughput of a network once its Attacking Case is calculated. The prediction model would be of assistance in the planning process of a network. This activity remains as our future work. It would be beneficial to use Attacking Case to predict the throughput as the Attacking Case metric could be calculated using simple procedure with the knowledge of node positions, transmission power, signal to interference ratio and radio propagation rather than using a discrete event network simulator. Network simulators demand simulator specific codes to be developed, multiple simulations to be executed, wait for the simulations to be completed, and output logs to be analyzed; only then one would have the knowledge on the expected throughput. As a result of this, the computation of the improved Attacking Case metric can be easily scaled up beyond the proposed simulation scenario.
