Interprocess Communications and Response Time

In modern applications, response time is crucial for software usability, and excessive waiting can lead to frustration and reduced productivity. However, the factors that contribute to poor response time are often misunderstood. While data structures and algorithms can impact performance in some cases, they are less likely to be the primary culprit in modern multitier enterprise applications. Instead, the number of remote interprocess communications (IPCs) that occur in response to a stimulus tends to have the most significant effect on response time.

Reducing the number of remote IPCs per stimulus can be achieved by optimizing the interface between processes, parallelizing interprocess communications where possible, or caching the results of previous IPCs to avoid future ones. In analyzing poorly performing applications, it is common to find ratios of thousands of IPCs to one stimulus. Therefore, when designing applications, it is important to consider the number of interprocess communications that will occur in response to each stimulus, as this is likely to have a more significant impact on response time than data structure choice or algorithm optimization. #SoftwarePerformanceTips