Cross-layer QoS cloud resource group scheduling optimization algorithm for power virtual hosts

In virtualized environments, QoS cloud resources in power virtual hosts exhibit dynamic elasticity and multi-dimensional constraints that lead to resource fragmentation during cross-layer scheduling. Resource fragments occupying scheduling channels under non-priority scheduling further increase delays. To address this, we propose a cross-layer QoS cloud resource group scheduling optimization algorithm for power virtual hosts. The physical layer provides support through the physical devices of the power virtual host structure, while the virtualization layer processes the resources of the physical layer through virtualization technology, dividing them into multiple virtual servers. To tackle the issues of resource fragmentation that can easily arise in cross-layer scheduling within a virtualized environment, we employ fuzzy clustering to perform descending grouping of cross-layer QoS cloud resources. The application layer combines the output arrangement results to construct a dynamic scheduling model for QoS cloud resources based on priority. To address the secondary fragmentation of resources within groups, we introduce a task priority scheduling method that perceives time slots to schedule resources according to priority, and implement a group admission control mechanism to optimize channel access during scheduling. We use a group arrival rate estimation algorithm to estimate the arrival rate of priority groups, and combine it with a dynamic coefficient to calculate the maximum number of available channels. Based on this, we determine the number of channels for resource scheduling and complete the scheduling optimization. Test results show that this algorithm can effectively achieve resource grouping under different resource quantities and categories, meeting the resource grouping scheduling requirements, with collection periods remain within 2 ms, and scheduling delay results all below 10ms; the maximum acceleration ratio for scheduling task execution is 5.92, satisfying the optimization requirements of cross-layer QoS cloud resource group scheduling in power virtual hosts.