AdaptScale: An adaptive data scaling controller for improving the multiple performance requirements in Clouds

作者：Shi, Yuliang; Dong, Mianxiong; Zhang, Wenbin; Liu, Lei^*; Zheng, Yongqing; Cui, Lizhen; Zhang, Junhua

来源：Future Generation Computer Systems-The International Journal of eScience, 2020, 105: 814-823.

DOI：10.1016/j.future.2017.08.034

• Multi-tenant
• Performance evaluation
• Self-adjust
• Data storage

Data scaling issue has become a bottleneck in multi-tenancy cloud environment. Fluctuated workloads bring challenges to current automatic data scaling strategies on meeting variable user performance requirements in a shared storage system. To this end, this paper develops an adaptive data scaling controller to meet multiple performance requirements in Clouds. The controller consists of three components: (1) a performance model, which determines whether the nodes are over-loaded: (2) a workload monitor and a predictor, which are responsible for collecting workload information and estimating the fluctuating trends, respectively: (3) a data scaling strategy generator, which enables the data scaling solution for over-loaded or under-loaded nodes. The numerical results show that the developed controller achieves the goal of automatic data scaling, which not only satisfies diversified performance requirements, but also reduces the execution time of MOVE operations with regards to system performance.

• 出版日期2020-4
• 单位山东大学