<i>FaaSLight</i> : General Application-Level Cold-Start Latency Optimization for Function-as-a-Service in Serverless Computing

Serverless computing is a popular cloud paradigm that frees developers from server management. Function-as-a-Service (FaaS) the most implementation of serverless computing, representing applications as event-driven and stateless functions. However, existing studies report functions FaaS severely suffer cold-start latency. In this paper, we propose an approach namely FaaSLight to accelerating cold start for through application-level optimization. We first conduct measurement study investigate possible root cause problem FaaS. The result shows application code loading latency significant overhead. Therefore, only indispensable can be adequate solution. Based on insight, identify related functionalities by constructing function-level call graph, separate other (i.e., optional code) applications. separated loaded demand avoid inaccurate identification causing failure. particular, key principle guiding design inherently general, i.e., platform - language-agnostic . practice, effectively applied developed in different programming languages (Python JavaScript), seamlessly deployed platforms such AWS Lambda Google Cloud Functions, without having modify underlying OSes or hypervisors, nor introducing any additional manual engineering efforts developers. evaluation results real-world show significantly reduce (up 78.95%, 28.78% average), thereby reducing As result, total response decreased up 42.05% (19.21% average). Compared with state-of-the-art, achieves 21.25 × improvement average