RTMC: an error control protocol for IP-based audio-visual multicast applications

Abstract

Transport of Audio Visual Datu is becoming an important application of the IP-based best-effoa service in the Internet. We present an error control protocol for provision of a real-time reliable multicast transport service. The protocol RTMC (Real-Time Multicast) uses Forward Error Correction (FEC) for protecting first transmissions (i.e., proactive error control) and retransmissions. Results of the pe~oimance evaluation show that RTMC error control imposes low latency, and is scalable to a large number of receivers. MotivationIn the past, a widespread believe has been that retransmission-based error control is not suitable for interactive audio-visual applications. While the development of retransmission-based error control mechanisms for continuous media (CM) applications has been a research topic for several years, it was only recently that this topic became of widespread interest, leading to a significant number of improvements. This includes a large number of proprietary protocols that are currently being introduced as part of CM Internet applications, and as Web browser plug-ins. This motivated our design goal of realizing an implementation concept of downloadable protocol modules. While the first generation of CM applications that deploy error control has been limited to simple point-to-point communication (unicast), recent developments also use multicast network services. As more and more end users get access to IP multicast services, error control schemes that allow the successful provision of multicast audiovisual services over the Internet are of significant economic importance. I Part of this work has been done while Georg Carle was In order to meet the reliability requirements of a real-time application, it is possible to use a network service that directly provides the required reliability, without additional error control mechanisms in the transport layer. This can be ensured by reservation of network resources, or by dimensioning the network in a way that the residual error probability is sufficiently small (over-engineering). In situations where a network service does not meet the reliability requirements of the application directly, additional error control mechanisms are required. By providing sufficient transport protocol processing capability with a low latency, it is possible to meet delay requirements of many audio-visual applications even after one or two retransmissions. This strategy potentially offers better utilization of network resources. It is particularly suitable for highly bursty traffic, as it allows to increase the load of intermediate systems up to a level in which losses are relatively frequent. As loss rates and loss characteristics may vary significantly, dimensioning error control parameters for proactive EEC and for retransmission needs to take into account the network scenario and the application scenario. IP best effort service provides no guarantees for loss rate, delay, and in-sequence delivery. This service model is based on the hope that all traffic sources are wel...