An On-Demand Rate-Adaptation Mechanism For IEEE 802.11 Networks Master Thesis

The current IEEE 802.11 standard mandates multiple transmission rates at the physical layer by employing different modulation and coding schemes. However, it does not specify any rate adaptation mechanism. It is left up to the researchers and vendors to implement adaptation algorithms that utilize the multi-rate capability. Rate adaptation algorithm is a critical component to the wireless system performance. The design of such algorithm is, however, not trivial due to the time-varying characteristics of the wireless channel (attenuation, collisions, interferences etc.). This has attracted the attention of researchers during the last few years. Previous work tends to select bit rates based on either frame loss statistics or physical layer (PHY) metrics, e.g., signal-to-noise ratio. While decisions in frame-based approaches are based on narrow information that limit their adaptability, the decisions in PHYbased approaches are more precise. However, the latter come with the overhead cost of transferring the channel information from the receiver to the transmitter. In this thesis we try to compromise between the channel adaptability and the cost of transferring channel information by signaling a minimized amount of information with respect to channel variations. This thesis presents a novel On-demand Feedback Rate Adaptation (OFRA) algorithm. The novelty of OFRA is that it allows receiver based adaptation through signaling channel information on a rate close to the channel coherence time. Hence, it eliminates the unnecessary overhead of transferring channel information at a fixed rate oblivious to the channel speed. In OFRA, the receiving node assesses the channel conditions by tracking the signal-to-noise ratio. Once it detects variations in the channel, it selects a new bit-rate and signals it back to the sending node. OFRA is to the best of our knowledge the first rate adaptation algorithm that can work in the absence of acknowledgments. This makes OFRA specially beneficial to nonacknowledged traffic that so far had to operate with a fixed bit-rate scheme. The throughput gains using OFRA stem from its ability to react fast in rapidly fluctuating channels and keeping the overhead low. Evaluation results obtained using NS-3 simulator show that OFRA consistently performs well in static as well as in mobile environments and outperforms ARF, Minstrel and Onoe.