Intertask message communication is another area where different operating systems show different timing characteristics. Most operating systems actually copy messages twice as they transfer them from task to task via a message queue. See Figure 6. The first copying is from the message-sender task to an operating system-owned “secret” area of RAM memory (implementing the “message queue”); and the second copying is from the operating system’s “secret” RAM area to the message-receiver task. Clearly this is non-deterministic in its timing, as these copying activities take longer as message length increases.
An approach that avoids this non-determinism and also accelerates performance, is to have the operating system copy a pointer to the message and deliver that pointer to the message-receiver task without moving the message contents at all. In order to avoid access collisions, the operating system then needs to go back to the message-sender task and obliterate its copy of the pointer to the message. For large messages, this eliminates the need for lengthy copying and eliminates non-determinism.
