计算机工程与设计
計算機工程與設計
계산궤공정여설계
COMPUTER ENGINEERING AND DESIGN
2015年
7期
1767-1770,1807
,共5页
云存储%多QoS约束%任务调度%存在矩阵%粒子群算法
雲存儲%多QoS約束%任務調度%存在矩陣%粒子群算法
운존저%다QoS약속%임무조도%존재구진%입자군산법
cloud storage%multiple QoS constrain%tasks scheduling%existing matrix%PSO
为研究云存储系统任务调度的问题,根据云存储系统的特点,前人已通过存在矩阵对PSO算法初始化与迭代更新进行约束,解决了PSO初始化以及迭代解对于云存储无意义的问题,使得PSO调度算法的迭代次数以及执行时间大幅降低,但其未充分考虑网络当前的状态以及网络服务质量问题。针对这一缺点,通过多QoS约束改进 PSO调度算法在QoS要求下的性能特征,使解更符合当前网络的状态以及用户对多QoS的需求。实验结果表明,虽然迭代次数和运行时间没有明显的变化,但是相对于现有算法在多QoS性能方面平均满足率仅有33%, QoS约束将该值改进至45.6%,满足多QoS的需求。
為研究雲存儲繫統任務調度的問題,根據雲存儲繫統的特點,前人已通過存在矩陣對PSO算法初始化與迭代更新進行約束,解決瞭PSO初始化以及迭代解對于雲存儲無意義的問題,使得PSO調度算法的迭代次數以及執行時間大幅降低,但其未充分攷慮網絡噹前的狀態以及網絡服務質量問題。針對這一缺點,通過多QoS約束改進 PSO調度算法在QoS要求下的性能特徵,使解更符閤噹前網絡的狀態以及用戶對多QoS的需求。實驗結果錶明,雖然迭代次數和運行時間沒有明顯的變化,但是相對于現有算法在多QoS性能方麵平均滿足率僅有33%, QoS約束將該值改進至45.6%,滿足多QoS的需求。
위연구운존저계통임무조도적문제,근거운존저계통적특점,전인이통과존재구진대PSO산법초시화여질대경신진행약속,해결료PSO초시화이급질대해대우운존저무의의적문제,사득PSO조도산법적질대차수이급집행시간대폭강저,단기미충분고필망락당전적상태이급망락복무질량문제。침대저일결점,통과다QoS약속개진 PSO조도산법재QoS요구하적성능특정,사해경부합당전망락적상태이급용호대다QoS적수구。실험결과표명,수연질대차수화운행시간몰유명현적변화,단시상대우현유산법재다QoS성능방면평균만족솔부유33%, QoS약속장해치개진지45.6%,만족다QoS적수구。
To study the task scheduling of cloud storage system ,according to the characteristics of the cloud storage system ,pre‐decessors have addressed the problem of PSO‐based arithmetic initialization and iteration by means of the exiting matrix ,w hich finds the solution to the cloud storage meaninglessness issue .Therefore ,the iterations and the execution time of PSO scheduling algorithm have been reduced significantly .Nevertheless ,it fails to take full account of the current state of the network and QoS . To solve this problem ,according to multiple QoS constrains ,the performance characteristics of the PSO scheduling algorithm under the requirement of QoS were improved ,and the solution could conform to the current network state and the user’s re‐quirement of multi‐QoS .The simulation results indicate that the iterations and the execution time do not change obviously .But as to the mere 33% fill rate of the existing multi‐QoS algorithm on average ,it can rise to 45.6% through QoS constrain .