The duplication of a data packet is handled by introducing unique identifiers. The sender tags each data packet with a unique identifier. The recipient records the identifiers of the data packets and discards packets whose identifiers have already been recorded. When packets are not reordered, the mechanism boils down to tagging packets with sequential numbers at the sender side and remembering the last number in each sequence at the recipient side.

Note that when the duplication of a data packet occurs due to the loss of an acknowledgement packet, the acknowledgement packet has to be sent even when the data packet is discarded, otherwise the sender would keep resending the data packet.