Classification and taxonomy of application-oriented workload prediction (IMAGE)
Caption
Two major characteristics of cloud applications affect their workload changes. Firstly, workload variability indicates the workloads of cloud applications exhibit significant volatility and pattern variability. The former refers to the fact that the inherent variance instability and noise perturbation of workloads, and the latter refers to the tendency of workloads to switch between different patterns over time. Secondly, workload heterogeneity represents the heterogeneity of four aspects: workload type, software architecture, runtime, and function type. Therefore, the design and implementation of workload prediction models must cater to these specific demands.
Credit
Tsinghua Science and Technology, Tsinghua University Press
Usage Restrictions
News organizations may use or redistribute this image, with proper attribution, as part of news coverage of this paper only.
License
Original content