A Task Execution Framework Based on Aging Indicator and SARIMI

The long-term running characteristics of spatiotemporal related tasks increase the risk of software aging of the basic supporting software of cloud platform, such as operating system, virtual machine and virtual machine monitor, and reduce the stability of platform services. In addition, software aging, as a process that accumulates over time, does not immediately cause system function failure. Moreover, the existing time series analysis methods do not give enough consideration to the seasonality of load characteristics, which makes it difficult to accurately determine the degree of software aging and the time of software rejuvenation. Therefore, this paper proposes a task execution method integrated with software rejuvenation means. Through the detection and evaluation of software aging, when the virtual machine aging reaches a certain extent, the rejuvenation module will be started to restore the service capacity of the virtual machine in time, so as to improve the task execution efficiency. And through the use of aging indicator and the software aging evaluation method based on Multiplicative Seasonal Autoregressive Integrated Moving Average (SARIMA), the current software aging situation and the periodicity of the system load are considered. At the same time, the dynamic trend of software aging is evaluated, and the software is rejuvenated near the optimal rejuvenation time point, which improves the applicability of resource aging evaluation and task execution efficiency in the cloud platform. Experimental results show that our method is effective in the aging of cloud platform software.