Performance Validation of Server-Less Computing Using Machine Learning

A promising paradigm for developing scalable and affordable cloud applications is serverless computing. However, because of their dynamic nature and intrinsic complexity, evaluating the performance of serverless platforms presents difficulties. With the use of machine learning techniques, this research offers a novel method for performance validating serverless computing environments. The machine learning models that can forecast performance under different workloads and setups are then trained using these metrics. Stakeholders can learn more about the behavior of serverless systems and adjust their applications accordingly by using this predictive modeling. The performance analysis carried out in the paper probably includes several variables, including throughput, scalability, reaction time, cost efficiency, and resource use. The authors aim to provide a relative performance of serverless edge computing systems by comparing these measures between various frameworks. Experiments are carried out on distinct serverless platforms, such as Kubeless, OpenFaas, Fission, and funX, to confirm the efficacy of the suggested methodology. The outcomes show how accurate and adaptable our prediction algorithms are at projecting performance indicators for various platforms and workload circumstances. Overall, by providing a methodical approach to performance assessment, this research advances serverless computing and helps cloud service providers and developers make more educated decisions.