In recent years, much work has been done on attempting to scale multicast data transmission to hundreds or thousands of receivers. There are, however, many situations where an application might involve transmission to just ten or twenty sites. The European Space Agency, for example, carry out regular multi-gigabyte bulk data transfers to a handful of destinations.Using multicast for this type of application can provide significant benefits including reduced load on the transmitter, an overall reduction in network traffic, and consequently shorter data transfer times.In this paper we take a fresh look at the problem of deploying reliable multicast. So far, there has been no convincing solution to achieve this. We present a simple hybrid solution which has not been proposed before. The approach taken is to combine unicast with multicast by modifying TCP to support multicast transfers, and run this modified TCP engine over UDP as a userspace transport protocol.Our goal is clear: reliable bulk data delivery to a moderate number of sites. Unlike some other multicast protocols, our work is complete: we have designed, implemented, deployed and evaluated a protocol which meets this goal.

In recent years, much work has been done on attempting to scale multicast data transmission to hundreds or thousands of receivers. There are, however, many situations where an application might involve transmission to just ten or twenty sites. Using multicast for this type of application can provide significant benefits including reduced load on the transmitter an overall reduction in network traffic, and consequently shorter data transfer times. In this project, we are investigating how partial or incomplete multicast can be exploited alongside reliable unicast to improve both speed and efficiency of data transfers while maintaining reliability. The approach taken is to combine unicast with multicast by modifying TCP to support multicast transfers, and run this modified TCP engine over UDP as a userspace transport protocol. We describe the work to date on the design and implementation, and provide experimental results from our tests across both local and wide area networks.