Approaches that are concerned with reliability issues in multicast protocols can be classified in sender-based and receiver-based.
In sender-based approaches, the sender maintains the state of all receivers to whom it transmits data and from whom it has to receive ACKs. For each packet, that a receiver obtains correctly, it sends a unicast ACK to the sender. If no ACK is received during a timeout period, the sender assumes that the packet is lost and retransmits it.
In contrast, in receiver-based approaches, each receiver informs the sender by NACKs (negative ACKs) about packets that are missing or received in error. And the sender retransmits these packets.
Both types of error control can be compared in multicast scenarios according to their ability to provide 100 percent reliability and their scalability with respect to the number of receivers.
In this application you are able to select the type of multicast error control and the scenario for a comparison (scalability and reliability). Try by yourself and find out which approach is guaranteed reliable and which scales well in large groups.
Please note: retransmissions in this application are always sent to the whole group!