Profiling Processor Usage in Web Applications: Categorization, Usage Patterns, and Experimental Analysis
摘要
The determination of the resource consumption patterns of modern web applications is increasingly critical, particularly given stringent non-functional requirements like low-latency user interactions. The increasing population of applications results in high competition for users; and user disatisfaction leads to engaging into some alternative system or app. Despite this, web development has traditionally prioritized functionality over performance, partly due to the perceived abundance of processing power of modern processors. This paper challenges that assumption and advocates for a conscious usage of profiling strategies in web systems that also focuses on the determination of the resource consumption of web application code. This paper explains the importance of determining execution time and its relation to the processor usage in the case of web code. We put forward a set of different strategies and mechanisms to determine the processor usage of web applications. We investigate a number of tools that are of interest for this purpose and provide its usage pattern. We experimentally assess these strategies through a set of experiments and provide a comparative analysis regarding their usability and obtained execution results.