科技通报
科技通報
과기통보
BULLETIN OF SCIENCE AND TECHNOLOGY
2014年
6期
158-160
,共3页
数据调度管理%多时隙%散布节点%大数据
數據調度管理%多時隙%散佈節點%大數據
수거조도관리%다시극%산포절점%대수거
data scheduling management%multi-time-slot%spread the node%big data
针对当前大数据任务的调度管理,传统的调度处理方法采用随机调度机制,从单个节点的性能出发,无法达到全局最优化的效果。提出基于改进多时隙散点算法的大数据任务调度管理方法,为了克服多时隙调度算法的缺点,运用多时隙调度算法时,引入了一个管理因子,对系统节点进行数据调度任务分配,根据分配的任务进行评价和估计,并且不断更新实时任务,追求最优性能,最终达到全局最优化。最后通过一组20节点数据进行测试实验,结果显示,采用改进的多时隙大数据任务调度管理算法,相对于传统调度方法,系统整体效率提高了平均约28%,具有很好的应用价值。
針對噹前大數據任務的調度管理,傳統的調度處理方法採用隨機調度機製,從單箇節點的性能齣髮,無法達到全跼最優化的效果。提齣基于改進多時隙散點算法的大數據任務調度管理方法,為瞭剋服多時隙調度算法的缺點,運用多時隙調度算法時,引入瞭一箇管理因子,對繫統節點進行數據調度任務分配,根據分配的任務進行評價和估計,併且不斷更新實時任務,追求最優性能,最終達到全跼最優化。最後通過一組20節點數據進行測試實驗,結果顯示,採用改進的多時隙大數據任務調度管理算法,相對于傳統調度方法,繫統整體效率提高瞭平均約28%,具有很好的應用價值。
침대당전대수거임무적조도관리,전통적조도처리방법채용수궤조도궤제,종단개절점적성능출발,무법체도전국최우화적효과。제출기우개진다시극산점산법적대수거임무조도관리방법,위료극복다시극조도산법적결점,운용다시극조도산법시,인입료일개관리인자,대계통절점진행수거조도임무분배,근거분배적임무진행평개화고계,병차불단경신실시임무,추구최우성능,최종체도전국최우화。최후통과일조20절점수거진행측시실험,결과현시,채용개진적다시극대수거임무조도관리산법,상대우전통조도방법,계통정체효솔제고료평균약28%,구유흔호적응용개치。
The traditional scheduling methods using random scheduling mechanism, starting from the performance of a sin-gle node, cannot reach the effect of the global optimization;Is proposed based on improved scatter algorithm of contention slot more big data scheduling management methods, first of all, in order to get rid of disadvantages of multiple time slot scheduling algorithm, using more time slot scheduling algorithms, introduces a management factor, the nodes of the system for data scheduling task allocation, and then evaluate the assigned tasks and estimates, and constantly updated real-time tasks, the pursuit of optimal performance, finally achieve global optimization. Finally through a set of 20 node data testing experiment, the results show that, with the improved much time slot scheduling management, compared with traditional scheduling method, the system overall efficiency is increased by an average of about 28%, it has the very good application value.