End to end performance monitoring approach

摘要

Performance monitoring is one of the most important aspects to ensure that software systems operate well. Performance administrators struggle to precisely monitor huge and complex systems with software and hardware components. Achieving this task in a simple and precise way is the goal of this research. To achieve this goal, we need to jump over internal system complexity, and must automate an intensive monitoring process. A simple and accurate approach was developed to monitor systems, all hardware and software components are managed as one abstract box, and then front-end services can be monitored by terms from end to end. The implemented case studies show and confirm that we can emulate manual performance monitoring by simulating the role of real subscriber exactly without any systematic restriction, hence complex algorithms, processes for status prediction and faults diagnosis aren’t farther needed in live system monitoring.