Measurement-based Call Admission Control Algorithm for Multi-Rate VoWLANs

Abstract

The combination of Voice over IP (VoIP) services and the widely deployed Wireless Local Area Networks (WLANs) has reached an enormous popularity in recent years due to the fact that it offers the possibility of voice communication with easy maintenance at a low cost. Unfortunately, achieving satisfactory service quality for the VoIP calls remains quite challenging due to the contention-based nature of the 802.11 Medium Access Control (MAC) that cannot guarantee the amount of resources required for successful voice transmission. This condition makes the capacity of the network a time varying parameter difficult to estimate. In order to avoid exceeding the network capacity with the admission of far too many voice sessions, which would result in the saturation of the network and the degradation of the voice performance, we design a Call Admission Control (CAC) algorithm. The CAC is in charge of accepting or rejecting new calls in the network in order to maintain a satisfactory Quality of Service (QoS) level for all the admitted voice sessions in the system. The proposed solution relies on a centralized CAC algorithm based on the real time monitoring of the activity in the network. The measurement of the events that occur in the network allows the CAC algorithm to accurately calculate the amount of airtime resources consumed by the active voice sessions. The fact that stations use different transmission rates and dynamically change them adds another source of randomness to the capacity estimation process and makes the calculation of the available resources even more challenging. We introduce an adjustment parameter based on the historical data transmission rate variations in order to make a conservative and educated admission decision. Consequently, the CAC decision is robust to potential rate changes in the network. Finally, our CAC solution includes the estimation of the impact that a new admission would make on the network and, once all these stages have been completed, the CAC algorithm checks the available resources and admits or rejects the requesting call accordingly. In this work, we conduct comprehensive ns-3 simulations of different working scenarios to evaluate the effectiveness of our proposed solution. In the simulation scenario where stationary voice stations use a fixed data transmission rate value, the proposed CAC algorithm performs very close to ideal, keeping the network resources efficiently utilized. Meanwhile, in the multi-rate environment studies (both the stationary and mobile node cases), our CAC algorithm manages to always maintain a satisfactory QoS level for all the admitted voice sessions even in the presence of Best Effort traffic. Eventually, we end up with a CAC algorithm capable of maintaining a satisfactory QoS level of the admitted call sessions regardless of the employed rate adaptation algorithms, while keeping an efficient use of the network resources. An improvement like this could make ubiquitous VoIP communication over WLANs possible.