Performance Analysis of Serverless Execution Environments

Serverless computing has introduced a novel Cloud delivery model that abstracts the infrastructure, platform, and execution layers, allowing users to focus on code development. With such abstraction, users lose insight into the factors that affect their serverless functions' execution performance. In this paper, we investigate the factors affecting serverless functions' execution performance. We examine the impact of container choices, language interpreter versions, and compilation options on a function's execution performance and execution consistency. The results of our investigation illustrate that seemingly trivial decisions can result in significant performance degradations for serverless functions' executions