Scalable reliable multicast using multiple multicast channels

We examine an approach for providing reliable, scalable multicast communica tion, involving the use of multiple multicast channels for reducing receive r processing costs and reducing network bandwidth consumption in a multicas t session, In this approach a single multicast channel is used for the orig inal transmission of packets. Retransmissions of packets are done on separa te multicast channels, which receivers dynamically join and leave. We first show that protocols using an infinite number of multicast channels incur m uch less processing overhead at the receivers compared to protocols that us e only a single multicast channel. This is due to the fact that receivers d o not receive retransmissions of packets they have already received correct ly, Next, we derive the number of unwanted redundant packets at a receiver due to using only a finite number of multicast channels, for a specific neg ative acknowledgment (NAK)-based protocol. We then explore the minimum numb er of multicast channels required to keep the cost of processing unwanted p ackets to a sufficiently low value (i.e., to achieve most of the benefit of using an infinite number of multicast channels), For an application consis ting of a single sender transmitting reliably to many receivers we find tha t only a small number of multicast channels are required for a Ride range o f system parameters, In the case of an application where all participants s imultaneously act as both senders and receivers a moderate number of multic ast channels is needed, Finally, we present two mechanisms for implementing multiple multicast channels, one using multiple TP multicast groups and th e other using additional router support for selective packet forwarding. We discuss the impact of both mechanisms on performance in terms of end-host (sender, receiver) and network resources. The approach of implementing mult iple multicast channels that uses additional router support reduces both en d-host processing costs and network bandwidth usage.