MOM vs. RPC: Communication Models for Distributed Applications

A n important aspect of designing a distributed application is the communication model you use to connect its various components. Some middleware solutions, 1 such as Web services, 2 offer two of the most important communication paradigms: asynchronous messaging and remote procedure call (RPC). Message-oriented middleware (MOM) lets a ser-vice's consumers physically and temporally decou-ple from the service providers (see www. webmethods.com and http://www-306.ibm.com/ software/integration/wmq/). Communication between service providers and their consumers is asynchronous, and they don't need to be available at the same time because they communicate by sending and receiving messages from designated message queues. In contrast, RPC is a synchronous method of requesting remote service execution. Consumers must suspend service execution until they receive a reply from the provider. MOM and RPC have advantages and disadvantages. MOM solutions tend to be more robust to failures than RPC, and they allow service re-questers to continue to process while service providers work on their requests. However, programming MOM-based applications is more cumbersome because distribution isn't as transparent to the programmer as with RPCs. In this column, I provide a quantitative framework you can use to compare MOM-and RPC-based solutions. Consider an application program component (APC) that requests a service from a service provider (SP). We assume that the APC could potentially execute some asynchronous code after sending a request for service but before receiving a reply from the SP; we call this time to execute the asynchronous code t async. With RPCs, the APC must wait until the RPC returns a result before it can execute the asynchronous code. With MOM, the APC and the SP exchange messages through a message queue broker (MQB), as Figure 1 shows. After sending a service request to the SP, the APC can execute its asynchronous code. The SP is constantly retrieving messages addressed to it from the MQB that contain requests for service execution. After completing this execution , the SP posts a message to the MQB addressed to the requesting APC, which retrieves the reply. Let's look at how to compute a request's execution time, from the APC's viewpoint, for RPC-and MOM-based communication with an SP (see Figure 2). We give an RPC-based request's execution time, T RPC , as T RPC = t N + t s + t N + t async = 2 ´ t N + t s + t async , (1) where t N is …