A HRFDC strategy based on dynamic classification of failed cloud tasks

Abstract

Abstract With the continuous development and improvement of cloud computing technology, the major computer giants have deployed their own cloud data center (CDC). At the same time, as user demands continue to expand, competition among cloud service providers is also intensifying. In order to continuously improve its own service quality and user satisfaction, cloud service providers adopting efficient and low-cost fault-tolerant strategy will improve the performance and profit of CDCs. However, the existing rescheduling strategy are mostly at the expense of the completion time of cloud task (CT) or increasing the compensation of cloud service providers, which ultimately leads to a decline in the profit of cloud service providers. More serious will affect the reputation and user experience of the enterprise. This paper systematically analyzes the performance loss caused by virtual machine (VM) failure and the rescheduling process of CDCs fault-tolerant strategy. At the same time, we established a dynamic classification rule of failed cloud task (FCT) according to the deadline for CTs. After that, a high-profit rescheduling fault-tolerant strategy for CDCs based on dynamic classification of FCTs (HRFDC) was proposed. This scheduling strategy maximizes the profitability of cloud service providers by increasing the failure repair rate of CDCs and reducing the compensation of cloud service providers. Finally, this strategy has been tested and verified, and its effect is due to the comparison algorithm.