基于滾動優(yōu)化的虛擬云中實時任務(wù)節(jié)能調(diào)度方法
發(fā)布時間：2018-12-14 07:49
【摘要】：目前,節(jié)能已成為云數(shù)據(jù)中心的研究熱點.建設(shè)節(jié)能的云數(shù)據(jù)中心不僅可以減少用電消耗,而且可以提高系統(tǒng)的可靠性.現(xiàn)有的云中心節(jié)能調(diào)度算法缺乏在任務(wù)調(diào)度級別的考慮,使得任務(wù)執(zhí)行效果受到較大影響.為此,首先給出了一種基于滾動優(yōu)化的實時任務(wù)調(diào)度器結(jié)構(gòu),然后詳細(xì)分析和構(gòu)建了任務(wù)能量消耗模型.在此基礎(chǔ)上提出了一種實時非周期任務(wù)節(jié)能調(diào)度算法EARH(energy-aware scheduling algorithm).EARH采用的滾動優(yōu)化策略能夠被拓展并集成其他節(jié)能調(diào)度算法.此外,提出了資源動態(tài)增加與縮減策略,用于在系統(tǒng)可調(diào)度性與節(jié)能兩方面進(jìn)行權(quán)衡.最后,通過大量的模擬實驗驗證了EARH的性能.與其他3種基準(zhǔn)算法相比,其實驗結(jié)果表明,EARH的調(diào)度質(zhì)量優(yōu)于其他算法,可有效提高系統(tǒng)性能.
[Abstract]:At present, energy saving has become the research hotspot of cloud data center. Building an energy-efficient cloud data center can not only reduce power consumption, but also improve the reliability of the system. The existing cloud center energy-saving scheduling algorithm lacks consideration of task scheduling level, which greatly affects the effect of task execution. Firstly, a real-time task scheduler structure based on rolling optimization is presented, and then the model of task energy consumption is analyzed and constructed in detail. On this basis, a real-time aperiodic task energy-saving scheduling algorithm EARH (rolling optimization strategy adopted by energy-aware scheduling algorithm). EARH) is proposed, which can be extended and integrated with other energy-saving scheduling algorithms. In addition, a dynamic resource increase and reduction strategy is proposed, which is used to balance the schedulability and energy saving of the system. Finally, the performance of EARH is verified by a large number of simulation experiments. Compared with the other three benchmark algorithms, the experimental results show that the scheduling quality of EARH is better than that of other algorithms, and the system performance can be improved effectively.
【作者單位】： 國防科學(xué)技術(shù)大學(xué)信息系統(tǒng)工程重點實驗室;
【基金】：教育部高等學(xué)校博士學(xué)科點專項科研基金(20134307110029) 湖南省自然科學(xué)基金(2015JJ3023) 西南電子電信技術(shù)研究室公開課題(2013001)
【分類號】：TP308
[Abstract]:At present, energy saving has become the research hotspot of cloud data center. Building an energy-efficient cloud data center can not only reduce power consumption, but also improve the reliability of the system. The existing cloud center energy-saving scheduling algorithm lacks consideration of task scheduling level, which greatly affects the effect of task execution. Firstly, a real-time task scheduler structure based on rolling optimization is presented, and then the model of task energy consumption is analyzed and constructed in detail. On this basis, a real-time aperiodic task energy-saving scheduling algorithm EARH (rolling optimization strategy adopted by energy-aware scheduling algorithm). EARH) is proposed, which can be extended and integrated with other energy-saving scheduling algorithms. In addition, a dynamic resource increase and reduction strategy is proposed, which is used to balance the schedulability and energy saving of the system. Finally, the performance of EARH is verified by a large number of simulation experiments. Compared with the other three benchmark algorithms, the experimental results show that the scheduling quality of EARH is better than that of other algorithms, and the system performance can be improved effectively.
【作者單位】： 國防科學(xué)技術(shù)大學(xué)信息系統(tǒng)工程重點實驗室;
【基金】：教育部高等學(xué)校博士學(xué)科點專項科研基金(20134307110029) 湖南省自然科學(xué)基金(2015JJ3023) 西南電子電信技術(shù)研究室公開課題(2013001)
【分類號】：TP308
【相似文獻(xiàn)】
相關(guān)期刊論文 前10條
1 聞育,吳鐵軍;基于蟻群算法的城域交通控制實時滾動優(yōu)化[J];控制與決策;2004年09期
2 王f，
本文編號：2378225
本文鏈接：http://www.sikaile.net/kejilunwen/jisuanjikexuelunwen/2378225.html
最近更新
教材專著
