A distributed measurement system for Internet delay estimation

Predicting the latency between hosts in the internet can play a significant role in the improvement of the performance of many services that use latency distances among hosts as a decision making input. Although, the information regarding the point to point delay among internet peers is required in many applications, such information is not easily available to the peers. Latency data acquisition requires a considerable amount of measurements to be performed at each node in order for that node to keep a record of its latency to all the other nodes. Moreover, the measurements need to be regularly repeated in order to be updated against the network dynamics where the latency values are frequently subject to change. This has motivated the use of techniques that alleviate the need for a large number of empirical measurements. Such techniques basically try to predict the entire latency matrix using a small set of latency measurements. Coordinate-based approaches are the most popular members among the family of latency prediction techniques. In these techniques, latency between each pair of nodes is modeled as the virtual distance among those nodes over a virtual system. This article proposes a new decentralized coordinate based approach to the problem of Internet delay measurement. Simulation results demonstrate the fact that the proposed system provides relatively accurate estimations.