Benchmarking parallelism in FaaS platforms

Serverless computing has seen a myriad of work exploring its potential. Some systems tackle Function-as-a-Service (FaaS) properties on automatic elasticity and scale to run highly-parallel jobs. However, they focus specific platforms convey that their ideas can be extrapolated any FaaS runtime. An important question arises: do all fit parallel computations? In this paper, we argue not them provide the necessary means host applications. To validate our hypothesis, create comparative framework categorize architectures four cloud offerings, emphasizing performance. We attest extend description with an empirical experiment consists in plotting deep detail evolution job each service. The analysis results evinces is inherently good for computations architectural differences across are decisive A key insight importance virtualization technologies scheduling approach platforms. Parallelism improves lighter proactive due finer resource allocation faster elasticity. This causes some like AWS IBM perform well computations, while others such as Azure present difficulties achieve required parallelism degree. Consequently, information paper becomes special interest help users choose most adequate infrastructure