王冠雅
摘 ?要: 針對云計算環(huán)境下滿足用戶服務質量(QoS)約束條件的在線服務性產品任務流分配問題,提出一種基于QoS約束的差分進化算法(QoS?DE算法),以便實現(xiàn)多目標優(yōu)化全局最優(yōu)問題。該算法首先構建了云計算環(huán)境下的QoS模型,并對在線服務性產品的工作流分配約束指標進行了分析。然后利用差分進化算法實現(xiàn)約束條件下的計算資源多目標優(yōu)化模型求解,并通過自適應的慣性權重調節(jié),提高了全局優(yōu)化能力。CloudSim云仿真平臺上的測試結果表明,相比經典Min?Min算法和QoS?GA算法,提出的QoS?DE算法能夠將任務合理分配到對應的節(jié)點,并在執(zhí)行時間、執(zhí)行費用等指標方面上表現(xiàn)出更好的性能。
關鍵詞: 云計算; 服務質量; 差分進化算法; 在線服務任務分配; 多目標優(yōu)化模型; QoS約束
中圖分類號: TN911.1?34; TP393 ? ? ? ? ? ? ? ? ?文獻標識碼: A ? ? ? ? ? ? ? ? ? ?文章編號: 1004?373X(2019)19?0132?03
Abstract: In order to solve the problem of task flow assignment of online service products that meet the user′s constraint conditions for quality of service (QoS) in cloud computing environment, a differential evolution algorithm based on QoS constraints (QoS?DE algorithm) is proposed, so as to achieve multi?objective global optimization. The QoS model in the cloud computing environment is constructed for the algorithm. The workflow allocation constraint indicators of online service products are analyzed. The differential evolution algorithm is used to solve the multi?objective optimization model of computational resources under constraint conditions, and the global optimization ability is improved by adaptive inertia weight adjustment. The test results on the CloudSim cloud simulation platform show that, in comparison with the classical Min?Min algorithm and QoS?GA algorithm, the proposed QoS?DE algorithm can reasonably assign tasks to the corresponding nodes, and has better performance in the aspects of execution time and cost indicators.
Keywords: Cloud computing; QoS; differential evolution algorithm; online service task allocation; multi?objective optimization model; QoS constaint
隨著互聯(lián)網時代信息與數(shù)據(jù)的快速增長,人們對存儲資源、帶寬和在線計算等網絡服務的需求越來越大。云計算作為一種新興的按需付費計算模式被提出來以便適應這些需求。云計算平臺能夠將數(shù)據(jù)中心的資源虛擬化,并充分利用網絡上閑置的資源為用戶提供服務。但是如何在復雜、動態(tài)、異構的環(huán)境中對云計算中的各種資源進行合理分配調度,并同時保證滿足用戶服務質量且系統(tǒng)負載均衡,是云計算的關鍵技術也是行業(yè)中一直關注的熱點方向。
傳統(tǒng)基于Web 服務技術的在線服務性產品工作流技術存在流程固定、柔韌性較差的問題,無法應對用戶的需求迅速增長、復雜性提高的新情況。不少動態(tài)資源任務調度算法被提出,例如,文獻[1]提出基于CSP的能耗高效云計算資源調度模型與算法,利用約束滿足問題對異構云數(shù)據(jù)中心的能耗優(yōu)化資源調度問題建模并求解,有效降低了云數(shù)據(jù)中心物理服務器的能耗。文獻[2]提出相對最小執(zhí)行時間方差的云計算任務調度算法min?variance,在CloudSim云仿真平臺測試的負載均衡和最早完成時間方面都達到較好的效果。大多數(shù)云計算資源調度問題可以視為一個NP全問題,即多目標優(yōu)化的問題。因此,文獻[3]提出一種基于遺傳算法的云計算資源調度策略,通過遺傳算法結合在平均負載約束條件下尋求全局負載最優(yōu)效果,提高了資源利用率。
本文提出基于QoS約束的差分進化算法(QoS?DE算法),能夠實現(xiàn)云平臺中在線服務性產品任務流分配問題,實現(xiàn)滿足用戶需求QoS約束條件(執(zhí)行成本最低和執(zhí)行時間最短)的計算資源,保證系統(tǒng)的負載均衡并為每個任務尋找合適的計算節(jié)點。通過仿真模擬驗證了QoS?DE算法在總執(zhí)行時間和總執(zhí)行費用這兩個指標上的性能表現(xiàn),優(yōu)于其他現(xiàn)有的方法。本文QoS約束的內容尚未包括云計算環(huán)境下的服務信譽等因素,考慮該因素在內的調度分配研究將會是下一步工作的重點。
參考文獻
[1] 林偉偉,劉波,朱良昌,等.基于CSP的能耗高效云計算資源調度模型與算法[J].通信學報,2013,34(12):33?41.
LIN Weiwei, LIU Bo, ZHU Liangchang, et al. Model and algorithm for energy?efficient cloud computing resource scheduling based on CSP [J]. Journal of communications, 2013, 34(12): 33?41.
[2] 李水泉,鄧泓.相對最小執(zhí)行時間方差的云計算任務調度算法[J].計算機技術與發(fā)展,2018(7):34?37.
LI Shuiquan, DENG Hong. Min?variance tasks scheduling algorithm for cloud computing system [J]. Computer technology and development, 2018(7): 34?37.
[3] 徐文忠,彭志平,左敬龍.基于遺傳算法的云計算資源調度策略研究[J].計算機測量與控制,2015,23(5):1653?1656.
XU Wenzhong, PENG Zhiping, ZUO Jinglong. Research on cloud computing resource scheduling strategy based on genetic algorithm [J]. Journal of computer measurement and control, 2015, 23(5): 1653?1656.
[4] YU W J, SHEN M, CHEN W N, et al. Differential evolution with two?level parameter adaptation [J]. IEEE transactions on cybernetics, 2017, 44(7): 1080?1099.
[5] WANG J, LIAO J, ZHOU Y, et al. Differential evolution enhanced with multiobjective sorting?based mutation operators [J]. IEEE transactions on cybernetics, 2017, 44(12): 2792?2805.
[6] ABUL KALAM AZAD M, FERNANDES E M G P. A modified differential evolution based solution technique for economic dispatch problems [J]. Journal of industrial & management optimization, 2017, 8(4): 1017?1038.
[7] ALI M, SIARRY P, PANT M. An efficient differential evolution based algorithm for solving multi?objective optimization problems [J]. European journal of operational research, 2018, 217(2): 404?416.
[8] HONG Z, JIANG H, BO L, et al. A framework for truthful online auctions in cloud computing with heterogeneous user demands [J]. IEEE transactions on computers, 2016, 65(3): 805?818.
[9] WEI Z, LIN Y, SHENG X, et al. Privacy preserving ranked multi?keyword search for multiple data owners in cloud compu?ting [J]. IEEE transactions on computers, 2016, 65(5): 1566?1577.
[10] WANG Y, LI J, WANG H H. Cluster and cloud computing framework for scientific metrology in flow control [J]. Cluster computing, 2017(1): 1?10.
[11] YOU C, HUANG K, CHAE H. Energy efficient mobile cloud computing powered by wireless energy transfer [J]. IEEE journal on selected areas in communications, 2016, 34(5): 1757?1771.
[12] DENG Ruilong, LU Rongxing, LAI Chengzhe, et al. Optimal workload allocation in fog?cloud computing toward balanced delay and power consumption [J]. IEEE internet of things journal, 2017, 3(6): 1171?1181.