Skip to Main Content
Despite the emergence of voice over IP and videoconferencing services, recent studies have shown that the Internet still may not provide acceptable quality for interactive streaming applications. Among the major parameters that affect quality of service are packet loss rates and loss burst sizes. To mitigate the negative impact of losses, forward error correction (FEC) techniques can be employed. These work by adding extra information to the data stream. So, while FEC can potentially reduce the perceived effects of packet losses, it increases bandwidth requirements which in turn may increase the loss rate. Therefore, it is important to choose, in real time, the proper FEC scheme to provide the best performance to the application. We evaluate the performance of parity-based FEC schemes using an analytical loss model. Then, we develop an adaptive mechanism for FEC selection using a predictive model. We present the results of simulation experiments using real Internet measurements.