Reliable Broadcast and Multicast in context of partitioning

We present two protocols, a distributed reliable broadcast and a distributed reliable multicast; these protocols are resistant to network partitions in asynchronous distributed systems. They are reliable in the sense that the receiver will eventually receive the message even if the receiver was partitioned away at any time. The protocols are aimed either for large scale or smaller networks. We show that our protocols are efficient in respect to the number of messages sent and complexity. We use an unreliable channel detector in conjunction with data forwarding and stability to achieve this goal. We present the pros and cons of broadcast over multicast and show that multicast is not well suited for broadcasting messages in the context of partitions. keywords: reliability, broadcast, multicast, partitionings, routing.