Desenvolvimento de um sistema distribuído com diferentes linguagens de programação para comparar o atraso do cold start em aplicações serverless

Abstract

Cloud computing has established itself as a powerful tool utilized in the market, and among the services offered, serverless computing— a software architecture deployed on a cloud computing platform that provides a hosting and execution environment—has stood out for its optimized consumption of computational resources, using them only when there is demand and releasing them when they are not needed. Additionally, it abstracts the complexity of infrastructure configurations for developers, allowing their efforts to focus on problem-solving. It is a recent concept and has gained notable popularity, with significant growth forecasts until 2025. However, it faces challenges, such as the cold start problem, which is the delay caused by the initialization of an instance in serverless, leading to processing drawbacks for applications. Thus, the objective of this study is to empirically examine how different programming languages influence the intensity of cold start in serverless applications. To this end, a distributed system was developed and implemented in different programming languages. Computational experiments were conducted on the developed system, and the generated data were analyzed and compared.