Cold Start Latency in Serverless Computing: Current Trends And Mitigation Techniques

Abstract

The newest cloud computing approach, serverless computing, makes it easier to design applications. Developers may avoid managing the servers by embracing and utilizing the Serverless Computing paradigm of today. The executable in this computational architecture is separate functions that are deployed one at a time on a Serverless platform with immediate per-request flexibility. The "Cold Starts" problem is usually brought on by such flexibility. The delay that happens when a runtime container is supplied to do the tasks is related to this issue. Other open-source and for-profit companies started adopting and providing this technology shortly after Amazon announced this computing architecture in 2014 with the AWS Lambda platform. Every Platform has a different way of handling cold starts. Over the past several years, great focus has been placed on issues affecting the cold start problem as well as the evaluation of each Platform's performance under load. This article offers a thorough summary of recent developments and cutting-edge research in reducing the cold start latency. Additionally, a number of sets of tests have been conducted to examine how the AWS Lambda base platform behaves with regard to the cold start latency.