A major advantage of using named pipes is that they provide a useful way to send one-line requests to an OpenEdge background session running a message handler procedure. Multiple users can send requests through the same named pipe and each request is removed from the pipe as it is received. In addition, the message handler procedure can loop indefinitely looking for input because it blocks (waits) until there is something to read. Finally, output through named pipes is more efficient than writing a complete response to an ordinary file, closing the file, and then informing the recipient that the results are available. The receiving process can read the result through a named pipe as soon as it is written.

A disadvantage of named pipes is that multiple processes cannot use a single named pipe to send or receive multi-line messages, unless you define a more complex protocol to control message interleaving. Also, although synchronizing named pipe input and output is helpful in some situations, it is a problem in others. For example, if the message handler procedure running in the background OpenEdge session starts returning results to an output named pipe, and for some reason the requestor is not ready to read the results, the message handler cannot move on to read the next request.