Exact Analysis of Rate Adaptation Algorithms in Wireless LANs

TL;DR

This paper provides an exact semi-Markovian analysis of ARF and AARF rate adaptation algorithms in wireless LANs, introduces a new variant PAARF, and compares their throughput performance under different channel conditions.

Contribution

It offers a novel exact analytical framework for ARF and AARF, introduces PAARF as a new algorithm, and compares their performance with insights on MAC overhead effects.

Findings

AARF and PAARF scale better at higher bit rates.

Neither ARF nor AARF consistently outperforms the other.

PAARF often performs close to the best in various regimes.

Abstract

Rate adaptation plays a key role in determining the performance of wireless LANs. In this paper, we introduce a semi-Markovian framework to analyze the performance of two of the most popular rate adaptation algorithms used in wireless LANs, namely Automatic Rate Fallback (ARF) and Adaptive Automatic Rate Fallback (AARF). Given our modeling assumptions, the analysis is exact and provides closed form expressions for the achievable throughput of ARF and AARF. We illustrate the benefit of our analysis by numerically comparing the throughput performance of ARF and AARF in two different channel regimes. The results show that neither of these algorithms consistently outperforms the other. We thus propose and analyze a new variant to AARF, called Persistent AARF (or PAARF), and show that it achieves a good compromise between the two algorithms, often performing close to the best algorithm in each of the studied regimes. The numerical results also shed light into the impact of MAC overhead on the performance of the three algorithms. In particular, they show that the more conservative strategies AARF and PAARF scale better as the bit rate increases.