A method for high speed interprocess communications comprises four steps.
Initially, first and second processes can be attached to a message buffer
in a shared region of user memory. In addition, each process can have a
corresponding message queue. In a preferred embodiment, the attaching
step comprises the step of attaching first and second processes to a
message buffer in a shared region of user memory exclusive of operating
system kernel space. Second, message data from the first process can be
accumulated in a location in the message buffer. Third, a memory offset
corresponding to the location in the message buffer can be placed in the
message queue of the second process. Finally, the accumulated data at the
location corresponding to the offset can be used in the second process.
Consequently, the accumulated message data is transferred from the first
process to the second process with minimal data transfer overhead.