Modeling and Analysis of Frame Aggregation in Unsaturated WLANs with Finite Buffer Stations

Frame aggregation is one of the several enhancements proposed by IEEE 802.11 Task Group n to improve channel utilization. In frame aggregation, more than one data frame is encapsulated to form an aggregate, and once an aggregate is formed, a station contends to access the medium to transmit the entire aggregate. We refer to the number of data frames encapsulated within an aggregated frame as aggregate size. We claim that a static assignment of aggregate size leads to the following performance trade-off: a small value might be insufficient to mitigate the transmission overheads, thereby nullifying the whole purpose of frame aggregation; whereas, a large value might affect the quality of service experienced by higher layers due to the extra wait time to build an aggregate. In this paper, we characterize this trade-off by studying the impact of aggregate size on metrics like frame latency and channel utilization. To estimate these metrics, we model the transmission queue of an 802.11n station as a bulk service queuing system. We study the impact of aggregate size over a wide range of operating conditions covering several traffic arrival rates from higher layers, service distribution and data frame sizes. Apart from validating the existence of above-mentioned performance trade-off, our results indicate that the choice of aggregate size not only depends on the traffic arrival rate, but also (more interestingly) on data frame sizes. This calls for a dynamic assignment of aggregate size.