SCARF: serverless cloud auto resource framework

Abstract

Serverless computing has revolutionized cloud computing by introducing automatic resource scaling based on demand. Nevertheless, addressing efficiency and performance challenges remains crucial for maximizing the benefits of serverless functions. This article introduces a novel framework for serverless computing that leverages machine learning technologies: the Serverless Cloud Auto Resource Framework (SCARF). SCARF represents a promising solution for predictive resource management in serverless computing environments. This thesis provides a comprehensive analysis of SCARF, including its architecture, functionality, and performance, through a rigorous validation process. The investigation reveals SCARF's ability to quickly adapt to the dynamic demands of serverless computing, with a significant reduction in prediction errors indicating its potential for precise resource allocation. By combining the flexibility of serverless computing with the power of machine learning, we aim to advance the state of the art in cloud resource management.