What is it about?
In this paper, we present an experimental design to evaluate the elasticity, a platform's ability to react to varying workloads, of the four leading serverless computing platforms. Further, we discuss changes in these serverless platforms' elasticity based on experimental results from late-2019 and mid-2020. Overall we show that all but one of the tested platforms can manage workloads with more than 100 requests per second without observable issues.
Featured Image
Photo by Taylor Vick on Unsplash
Why is it important?
Serverless cloud computing platforms, specifically, Function-as-a-Service platforms, users a simplified programming model, a pay-per-use billing model, and an elastic computing platform. This paper evaluated the promise of elasticity and showed that not all cloud providers could manage an equal level of elasticity. Thus, users of these platforms should be aware of these differences and evaluate if their workload fits these platforms' real capability.
Perspectives
I hope developers using FaaS platforms are aware of the differences between the well-known cloud providers. With our model and toolkit, developers can check if their use-case is a good fit for a specific platform, thus, avoiding performance issues down the road.
Sebastian Werner
Technische Universitat Berlin
Read the Original
This page is a summary of: All but one, ACM SIGAPP Applied Computing Review, September 2020, ACM (Association for Computing Machinery),
DOI: 10.1145/3429204.3429205.
You can read the full text:
Resources
Contributors
The following have contributed to this page
