With the advent of overlay and peer-to-peer networks, Grid computing, and CDNs, network performance prediction becomes an essential task. Predicting the throughput of large TCP transfers, in particular, has attracted much attention.

In this work, we focus on the design, empirical evaluation, and analysis of TCP throughput predictors for a broad class of applications. We first classify TCP throughput prediction techniques into two categories: Formula-Based (FB) and History-Based (HB). Within each class, we develop representative prediction algorithms, which we then evaluate empirically over the RON testbed. FB prediction relies on mathematical models that express the TCP throughput as a function of the characteristics of the network path (e.g., RTT, loss rate, available bandwidth).

FB prediction does not rely on previous TCP transfers in the given path, and it can be performed with non-intrusive network measurements. We show, however, that the FB method is accurate only if the TCP transfer is window-limited to the point that it does not saturate the underlying path, and explain the main causes of the prediction errors. HB techniques predict the throughput of TCP flows from a time series of previous TCP throughput measurements on the same path, when such a history is available. We show that even simple HB predictors, such as Moving Average and Holt-Winters, using a history of limited and sporadic samples, can be quite accurate. On the negative side, HB predictors are highly path-dependent. Using simple queueing models, we explain the cause of such path dependencies based on two key factors: the load on the path, and the degree of statistical multiplexing.