面向嵌入式系统的多特征融合任务调度方法OACSTPCD
A Multi-feature Fusion Task Scheduling Method for Embedded Systems
随着容器技术在嵌入式系统中逐渐被广泛采用,面向嵌入式容器的任务调度就成为了一项热点研究内容.虽然云服务器环境下的调度方法已经比较成熟,拥有像动态优先级抢占式调度、负载均衡调度等丰富的调度算法,但由于嵌入式系统具有与传统云服务器不同的特点,嵌入式系统通常需要按照特定的应用场景和需求进行定制化的操作系统和驱动程序开发,以适应不同的硬件组合和资源限制,使得传统云服务器环境下的调度方法无法满足嵌入式系统下的任务调度的实时性和灵活性要求.为此,提出一种多特征融合的任务调度方法.该方法在实时调度任务依赖和任务特征分析的基础上,综合考虑能耗价值、时间价值和负载价值三个特征因素,以动态抢占方式实现实时任务调度.通过测试方法的设计证明,该方法可以有效帮助嵌入式系统进行容器资源配额,还为生产线运作带来了稳定性和紧急情况下的危机处理能力.
With the gradual adoption of container technology in embedded systems,task scheduling for embedded containers has become a hot research topic.Although the scheduling method in the cloud server environment is relatively mature,with rich scheduling algorithms such as dynamic priority preemptive scheduling,load balancing scheduling,etc.,the embedded system has different characteristics from the traditional cloud server,the embedded system usually needs to follow the customized operating system and driver development for specific application scenarios and requirements to adapt to different hardware combinations and resource constraints,making the scheduling method in the traditional cloud server environment unable to meet the real-time and flexibility of task scheduling in embedded systems sexual demands.For this reason,a multi-feature fusion task scheduling method is proposed.Based on the analysis of real-time scheduling task dependencies and task characteristics,the proposed method comprehensively considers the three characteristic factors of energy consumption value,time value and load value,and realizes real-time task scheduling in a dynamic preemptive manner.The design of the test method proves that the proposed method can effectively help the embedded system to perform container resource quotas,and also bring stability and emergency crisis handling capabilities to the operation of the production line.
田今晨;赵卓峰
北方工业大学 信息学院,北京 100144||大规模流数据集成与分析技术北京市重点实验室(北方工业大学),北京 100144
计算机与自动化
嵌入式容器技术资源配额任务调度价值
embeddedcontainer technologyresource quotatask schedulingvalue
《计算机技术与发展》 2024 (007)
1-8 / 8
北京市自然科学基金项目(4202021)
评论