One of the goals of any software design is to produce software which behaves predictably and reproducibly under a given stimulus. Many object-oriented and message-oriented design paradigms suffer because they introduce a degree of unpredictability and randomness into a software product. If a message is exchanged with a mailbox scheme and the sender and the receiver are never brought into synchronization it will be exceedingly difficult to replicate or predict all possible states that two process system can be in. Much depends on the environment and timing considerations in which these two processes find themselves. These modules may work fine in one environment and suffer sporadic failures in another. This leads to increased testing and maintenance costs and a poorer software product. Often the blame is incorrectly directed at the multiprocess design paradigm. How often have we heard it said that client/server software design does not handle complexity well. The original designers of Send/Receive/Reply synchronized messaging believed that the answer lies in forcing a state-machine-like synchronization to occur on each message pass.
It works like this: a sending process composes a message and arranges to send it to a receiving process. While waiting for a reply the sending process is blocked. The receiving process, on the other hand, is held blocked until a message arrives destined for it. It unblocks, reads the message, processes its contents and then replies to the sender. This reply then unblocks the sender and the two processes are free to operate independently once again.
Many have argued that this blocking/forced synchronization introduces unnecessary complexity into a message exchange, but when properly applied it achieves exactly the opposite effect. By forcing synchronization to occur at each message pass, one finds that our multi-process applications now start to behave in a very predictable and reproducible manner. Gone are the timing and environmental effects that often plague nonsynchronized message passing schemes. When dealing with huge complex applications, this represents a huge strategic advantage. As an added bonus, since the sender is blocked during message transmission and is explicitly unblocked by the receiver process with the reply, it is very simple to arrange to transport these messages over a variety of media (including some which are “slow”). The messages could be exchanged via shared memory if the processes were on the same processor, or they could be exchanged via the Internet if the processes were physically separated or on a serial line in a dial up situation. While the performance of the collective of processes would obviously be affected by the message transmission speed, the predictability and reproducibility of that performance would not be.
The importance of predictability of software cannot be overstated when it comes to software testing. Nothing makes software QA people wish for a career change more than an application which exhibits unpredictable and unreproducible behavior.
To help promote this software IC paradigm in the Linux community we started the SIMPL open-source project (www.holoweb.net/~simpl). SIMPL enabled processes are Linux processes with all the features and protections that affords. SIMPL enabled processes are able to exchange tokenized binary messages using a blocking Send/Receive/Reply messaging scheme. In short SIMPL processes have the makings of some great software ICs.
Before going into more detail on how these software ICs might be built, it is worth emphasizing the advantages of this model.
In principle, SIMPL processes can be written in any language. While much of the code on the SIMPL web site is still written in C, there is no reason why a C++ or JAVA SIMPL process could not be created and talk transparently to another SIMPL process written in Tcl/Tk or Python.
The language used to write the software IC itself in no way dictates the language of another software IC with which interaction takes place. Furthermore, a given SIMPL process has no way of discovering what language was used to construct another IC with which it is exchanging a message.
A SIMPL software IC has no way of discovering or knowing the physical location of its exchange partner. This means that the same binary image can be tested with local message exchanges and then deployed with remote message exchanges. Overall collective application performance would differ but the individual software IC would not need to change in any way. In may instances even a recompile is unnecessary.
A SIMPL software IC has no way of discovering the internal algorithm of an exchange partner. This means that test stubs can be created which completely simulate and replicate the environment in which a given software IC finds itself. In particular, error conditions which would be costly or difficult to reproduce in the full system can readily be simulated in a test environment. These SIMPL software ICs can be rigorously tested before being deployed in the real-world application.
SIMPL software ICs lend themselves well to projects with multiple developers. The implementation details of a SIMPL enabled process cannot affect any interacting process, provided that implementation conforms to the agreed upon message API. While a poor IC implementation will certainly adversely affect the overall application performance, the application will still operate. Once a poor algorithm has been identified it can be worked on in isolation from the real application and substituted once tested without even recompiling the adjacent ICs.