Elastic Container Scheduling for Stochastically Arrived Workflows in Cloud and Edge Computing

Dong Wen,Lixin Zhu,Jian Xu,Zhicheng Cai
DOI: https://doi.org/10.1007/978-981-19-4546-5_4
2022-01-01
Abstract:Many intelligent applications of Cloud and Edge can be modeled to be stochastically arrived workflows. Elastically provisioning resources such as containers to these applications is crucial to minimize resource costs while fulfilling deadlines. Most of existing workflow scheduling algorithms are either not for randomly arrived workflows from users of Edge Computing or only consider workflows in pure Cloud Computing. In this article, an elastic resource scheduling method, which integrates loosely coupled workflow scheduling with resource auto-scaling, is developed for stochastically arrived workflows in Cloud and Edge to minimize long-term operating costs while meeting deadline constraints. Experimental results illustrate that our approach achieves a lower overall cost and a higher success ratio compared with the state-of-art algorithm.
What problem does this paper attempt to address?