面向电力大数据的广域计算任务流规划调度技术研究  

Research on Task Flow Planning and Scheduling in Wide Area Computing for Power Big Data

作　　者：朱力鹏 饶玮 裘洪彬 吴舜 靳丹[3] ZHU Li-peng RAO Wei QIU Hong-bin WU Shun JIN Dan(Global Energy lnterconnection Research Institute, Nanjing 210003, China State Grid Jibei Electric Power Company, Beijing 100053, China State Grid Gansu Electric Power Company, Lanzhou 730046, China)

机构地区：[1]全球能源互联网研究院,江苏南京210003 [2]国网冀北电力公司,北京100053 [3]国网甘肃省电力公司,甘肃兰州730046

出　　处：《电力信息与通信技术》2017年第3期62-66,共5页Electric Power Information and Communication Technology

摘　　要：随着电力大数据的深化应用,对基于多数据中心的任务规划、调度、计算能力的均衡和计算节点的效率提出了更高的要求,文章详细阐述了电力大数据广域计算任务流规划与调度问题,探讨了适用于多数据中心的广域计算任务流处理框架,介绍了顾及业务逻辑关系的最优粒度任务分解方法,分析了面向电力大数据的广域计算任务流规划策略,提出面向电力大数据的分布式任务协同调度方法,为电力大数据广域并发计算任务流的合理规划与协同调度提供技术参考,综合提升电力大数据广域并发计算任务流的处理效率。With the deepening application of power big data, higher requirement based on multi data center mode is put forward, including the task scheduling, computing power equalization and computing nodes efficiency. In this paper, the task scheduling and scheduling problem of big data wide area computing is discussed in detail. And the power big data area calculation problem of task flow planning wide area computing task flow processing for multi data center and scheduling is described and the framework of is discussed. A distributed task scheduling method based on power big data is proposed. It provides a technical reference for the reasonable planning and collaborative scheduling of the wide area data concurrent computing tasks, and comprehensively improves the processing efficiency of big data concurrent computing task flow.

关 键 词：电网 大数据 数据中心 任务流 规划策略 协同调度 

分 类 号：TP319[自动化与计算机技术—计算机软件与理论]