A simple but effective optimisation for a popular level‐5 protocol called Java's Remote Method Protocol, which is used in a distribution model named RMI (Java's Remote Method Invocation) is presented. The protocol has been enhanced with a simple multiplexing mechanism which offers the possibility of transferring several parallel request–response interactions without opening new transfer control protocol/Internet protocol connections. The temporal overhead and the advantages stemmed from the approach in terms of response‐time are explored in a simple but realistic evaluation scenario: a local area network (LAN) (switched‐Ethernet) infrastructure under the small packet size condition (i.e. when the remote invocation does not have packet fragmentation). Presented results show that the proposed extensions are useful in the LAN scenario under the non‐fragmentation assumption in applications that do not require packet fragmentation.