A task container scheduling algorithm for spaceborne virtualization platform

一种用于星载虚拟化平台的任务容器调度算法

¹ School of Computer Science, Northwestern Polytechnical University, Xi'an 710072, China
² Xi'an Institute of Microelectronics Technology, Xi'an 710065, China

Received: 23 February 2023

Abstract

With the help of lightweight virtualization technology such as containers, the spaceborne virtualization platform encapsulates computing tasks into containers to form task containers, so as to achieve efficient utilization of resources. However, the platform's task container scheduling problem is a difficult problem that needs to be solved urgently. In this paper, we aim at this problem by establishing a multi-pass scheduling model for separable container tasks based on non-blocking communication mode. On the basis of this model, we propose a new scheduling algorithm, aiming at determining the optimal processor scheduling sequence and scheduling times. The algorithm combines the concept of divisible task container and multi-pass scheduling. By decomposing the task into executable subtasks, and performing task allocation and processor scheduling in multiple scheduling stages, it can optimize the scheduling order to improve the overall processing efficiency. This algorithm is an improved genetic algorithm that adds the optimization strategy of subpopulation isolation to the traditional genetic algorithm. Its core idea is to improve the performance and effect of the genetic algorithm by introducing the population division strategy into the algorithm process. We verify the effectiveness and convergence of the algorithm through experiments, and the experimental results show that the algorithm makes the task have less completion time.

摘要

星载虚拟化平台借助容器等轻量级虚拟化技术, 将计算任务封装到容器中形成任务容器, 从而实现资源的高效利用。然而, 该平台的任务容器调度问题是一个亟需解决的难题。针对这一问题建立了一个基于非阻塞通信模式的可分容器任务多趟调度模型。在该基础上, 提出了一种新的调度算法, 旨在确定最佳的处理机调度顺序和调度趟数。该算法结合可分任务容器和多趟调度的概念, 通过将任务分解为可执行的子任务, 并在多个调度阶段中进行任务分配和处理机调度, 从而优化调度顺序, 提高整体处理效率。该算法是一种改进的遗传算法, 在传统遗传算法的基础上添加了子种群隔离的优化策略, 其核心思想是将种群划分策略引入算法过程, 从而改善遗传算法的性能和效果。通过实验验证了该算法的有效性和收敛性, 结果表明, 该算法缩短了任务完成时间。