Estimating link capacity in a wireless network is a complex task because theavailable capacity at a link is a function of not only the current arrival rateat that link, but also of the arrival rate at links which interfere with thatlink as well as of the nature of interference between these links. Models whichaccurately characterize this dependence are either too computationally complexto be useful or lack accuracy. Further, they have a high implementationoverhead and make restrictive assumptions, which makes them inapplicable toreal networks. In this paper, we propose CapEst, a general, simple yet accurate,measurement-based approach to estimating link capacity in a wireless network.To be computationally light, CapEst allows inaccuracy in estimation; however,using measurements, it can correct this inaccuracy in an iterative fashion andconverge to the correct estimate. Our evaluation shows that CapEst alwaysconverged to within 5% of the correct value in less than 18 iterations. CapEstis model-independent, hence, is applicable to any MAC/PHY layer and works withauto-rate adaptation. Moreover, it has a low implementation overhead, can beused with any application which requires an estimate of residual capacity on awireless link and can be implemented completely at the network layer withoutany support from the underlying chipset.
展开▼
