Comparison of End-to-End Loss Differentiation Algorithms in Multi-hop Wireless Networks

When TCP operates in wireless networks, it suffers from severe performance degradation. This is because TCP reacts to wireless packet losses by unnecessarily slowing down its sending rate. Although several loss differentiation schemes are proposed to avoid such performance reduction, their accuracies were not evaluated in multi-hop wireless networks because these schemes are designed for only the last-hop wireless networks. In this paper, not only we observe how the accuracies of these schemes vary in multi-hop wireless networks, but also we suggest a new end-to-end loss differentiation scheme. Our scheme estimates the rate of queue usage using information available to TCP. If the estimated queue usage is larger than 50% when a packet is lost, our scheme diagnoses the packet loss as congestion losses. Otherwise, it diagnoses the packet loss as wireless losses. Through extensive simulations, we show that the accuracies of previous schemes tend to decrease as the number of hops increases, and also show that our scheme has the highest accuracy in multi-hop wireless networks.