1 School of Electrical and Information Engineering, Beijing University of Civil Engineering and Architecture, Beijing 102616, China
2 Research on Intelligent Processing Method of Building Big Data Beijing Key Laboratory (Beijing University of Civil Engineering and Architecture), Beijing 102616, China
Show less
文章历史+
Received
Published
Revised
2024-11-07
2028-07-28
2025-11-03
Online Date
2026-09-20
PDF (4136K)
摘要
云计算系统需要进行准确的主动式资源分配以实现高质量的云服务和高效的云资源利用,而恰当的主动式资源分配需要对工作负载进行准确的预测。现有的很多适用于云计算系统的工作负载预测方法要么预测准确性有限,要么因开销过大而缺乏实用性。针对上述问题提出了基于空洞卷积神经网络的虚拟机工作负载预测算法(virtual machine workload prediction algorithm based on dilated convolutional neural network,VMWPD),该算法采用了基于空洞卷积神经网络的预测模型,应用了预测模型预测工作负载的变化量而不是直接预测工作负载的机制,并拥有实时预测和实时训练的能力,在预测准确度和开销之间取得了较好的平衡。评估实验结果表明,VMWPD相较于基于长短时记忆模型(long short-term memory,LSTM)的工作负载预测算法准确度提高了32.45%,且时间开销降低了45.10%。可见,本文方法在保证一定精度的情况下能大幅降低开销。
Abstract
Cloud computing systems require accurate proactive resource allocation to achieve high quality cloud services and efficient cloud resource utilization, and proper proactive resource allocation requires accurate workload prediction. Many existing workload prediction methods applicable to cloud computing systems either have limited prediction accuracy or lack practicality due to excessive overhead. A virtual machine workload prediction algorithm based on dilated convolutional neural network (VMWPD) was proposed to address the above problems. The algorithm adopted a prediction model based on dilated convolutional neural network, applied the mechanism that the prediction model predicts the amount of change in the workload instead of directly predicting the workload, and possessed the ability of real-time prediction and real-time training, which striked a better balance between prediction accuracy and overhead. Evaluation experimental results show that VMWPD improves the accuracy of workload prediction algorithm by 32.45% and reduces the time overhead by 45.10% compared to long short-term memory (LSTM) based workload prediction algorithm. As can be seen, the proposed method significantly reduces the cost while maintaining a certain level of accuracy.
LIU Hao-yu, TIAN Le, GUO Mao-zu.
Virtual Machine Workload Prediction Algorithm Based on Dilated Convolutional Neural Network[J].
Science Technology and Engineering, 2026, 26(3): 1128-1134 DOI:10.12404/j.issn.1671-1815.2408305
基于回归的方法通常采用统计学的方法建立预测值与已知的历史值之间的联系,然后利用历史值与回归模型计算出预测值。Nisar等[11]利用差分整合滑动平均自回归模型(autoregressive integrated moving average model,ARIMA)分析数据中心的特定工作负载的时间序列特征,然后选取不同的数学模型拟合工作负载序列以进行预测。Feng等[12]提出了一种自适应滑动窗口算法,算法有能力考虑工作负载的趋势相关性、时间相关性和随机波动性,还开发了一种用于模型集成的多类回归算法,在云应用预测中实现了更佳的准确性。Bi等[13]针对云数据中心的工作负载预测问题提出了一种新颖的预测机制,使用Savitzky-Golay滤波器对描述工作负载的时间序列数据进行平滑处理,并采用小波分解对数据进行划分,还集成了ARIMA,该方法能够更准确地预测时间间隔内的工作负载。许多基于回归的模型直接或间接地假设时间序列数据是平稳的,而且一些统计学方法难以很好地对时间序列中的非线性关系进行建模,这限制了此类方法的预测精度与适用范围。
选取基于LSTM的工作负载预测方法[28]和基于双向长短时记忆模型(bidirectional long short-term memory,BLSTM)的工作负载预测方法[29]作为对比算法。实现了可以在本实验环境下运行的这些对比算法,为了实验的公平性,尽可能让所有参与实验的算法中需要训练的参数个数接近,并根据实验环境对所有参与实验的算法进行了调优。
4.3 评价指标
选取归一化均方根误差(normalized root mean squared error,NRMSE)和归一化平均绝对误差(normalized mean absolute error,NMAE)这两个评价指标用于评估预测算法的预测性能,这两个评价指标的计算方式为
FuY, LiC, YuF R, et al. A survey of blockchain and intelligent networking for the metaverse[J]. IEEE Internet of Things Journal, 2023, 10(4): 3587-3610.
[2]
AsghairA, SohrabiM K. Server placement in mobile cloud computing: a comprehensive survey for edge computing, fog computing and cloudlet[J]. Computer Science Review, 2024, 51: 100616.
WangTing, ShaoPeng, LiGuangquan, et al. Improved particle swarm optimization algorithm for cloud computing task scheduling[J]. Science Technology and Engineering, 2023, 23(29): 12594-12603.
[5]
BiswasD, JahanS, SahaS, et al. A succinct state-of-the-art survey on green cloud computing: challenges, strategies, and future directions[J]. Sustainable Computing: Informatics and Systems, 2024, 44: 101036.
[6]
ZhangL, XieY, JinM, et al. A novel hybrid model for docker container workload prediction[J]. IEEE Transactions on Network and Service Management, 2023, 20(3): 2726-2743.
[7]
WaliaG K, KumarM, GillS S. AI-empowered fog/edge resource management for IoT applications: a comprehensive review, research challenges and future perspectives[J]. IEEE Communications Surveys & Tutorials, 2024, 26(1): 619-669.
[8]
SeshadirK, SindhuK, BhattuS N, et al. Design and evaluation of a hierarchical characterization and adaptive prediction model for cloud workloads[J]. IEEE Transactions on Cloud Computing, 2024, 12(2): 712-724.
XuWan, XiaRuidong, LiuDongting, et al. Task scheduling algorithm in ship edge computing based on improved NSGA-Ⅲ[J]. Science Technology and Engineering, 2024, 24(33): 14306-14315.
ZhaoChanchan, GuoXiaomin, HaiXiaowei, et al. Joint optimization strategy of task offloading and resource allocation for cache-assisted mobile edge computing[J]. Science Technology and Engineering, 2023, 23(9): 3812-3819.
[13]
DoganiJ, NamvarR, KhunjushF. Auto-scaling techniques in container-based cloud and edge/fog computing: taxonomy and survey[J]. Computer Communications, 2023, 209(1): 120-150.
[14]
NisarF, BaseerS, KhanA. Survey on ARMIA model workloads in a datacenter with respect to cloud architecture[C]// Proceedings of the International Symposium on Recent Advances in Electrical Engineering. Islamabad: IEEE, 2019: 1-4.
[15]
FengB, DingZ, JiangC. FAST: a forecasting model with adaptive sliding window and time locality integration for dynamic cloud workloads[J]. IEEE Transactions on Services Computing, 2023, 16(2): 1184-1197.
[16]
BiJ, YuanH, LiS, et al. ARIMA-based and multiapplication workload prediction with wavelet decomposition and savitzky-golay filter in clouds[J]. IEEE Transactions on Systems, Man, and Cybernetics: Systems, 2024, 54(4): 2459-2506.
[17]
LiJ, YaoJ, XiaoD, et al. EvoGWP: predicting long-term changes in cloud workloads using deep graph-evolution learning[J]. IEEE Transactions on Parallel and Distributed Systems, 2024, 35(3): 499-516.
[18]
KimY, SongS, KooB, et al. Enhancing long-term cloud workload forecasting framework: anomaly handling and ensemble learning in multivariate time series[J]. IEEE Transactions on Cloud Computing, 2024, 12(2): 789-799.
[19]
ParkS, BahnH. Combining genetic algorithms and bayesian neural networks for resource usage prediction in multi-tenant container environments[J]. Cluster Computing, 2025, 28: 111.
[20]
YuF, KoltunV. Multi-scale context aggregation by dilated convolutions[C]// Proceedings of the International Conference on Learning Representations. San Juan: OpenReview. net, 2016: 1-8.
[21]
KhannaP, NarayanA. Light weight dilated CNN for time series classification and prediction[C]// Proceedings of the IEEE International Conference on Systems, Man, and Cybernetics. Toronto: IEEE, 2020: 2179-2183.
ChenJili, LiHaijun, XieXiaolan. Short-term cloud resource prediction model based on temporal convolution and long short-term memory[J]. Science Technology and Engineering, 2025, 25(7): 2856-2864.
[24]
BiJ, MaH, YuanH, et al. Accurate prediction of workloads and resources with multi-head attention and hybrid LSTM for cloud data centers[J]. IEEE Transactions on Sustainable Computing, 2023, 8(3): 375-384.
[25]
GaoZ, YangL, DaiY. Large-scale computation offloading using a multi-agent reinforcement learning in heterogeneous multi-access edge computing[J]. IEEE Transactions on Mobile Computing, 2023, 22(6): 3425-3443.
[26]
ShenS, HanY, WangX, et al. Collaborative learning-based scheduling for kubernetes-oriented edge-cloud network[J]. IEEE/ACM Transactions on Networking, 2023, 31(6): 2950-2964.
[27]
TuliS, PoojaraS R, SriramaS N, et al. COSCO: Container orchestration using co-simulation and gradient based optimization for fog computing environments[J]. IEEE Transactions on Parallel and Distributed Systems, 2022, 33(1): 101-116.
LiZhaoxi, LiuHongyan. Combining global and sequential patterns for multivariate time series forecasting[J]. Chinese Journal of Computers, 2023, 46(1): 71-84.
[30]
ShenS, BeekV V, IosupA. Statistical characterization of business-critical workloads hosted in cloud datacenters[C]// Proceedings of the IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing. Shenzhen: IEEE, 2015: 465-474.
[31]
MalekiE F, MashayekhyL, NabavinejadS M. Mobility-aware computation offloading in edge computing using machine learning[J]. IEEE Transactions on Mobile Computing, 2023, 22(1): 328-340.
[32]
LiX, ZhangS, HuangY, et al. Towards timely video analytics services at the network edge[J]. IEEE Transactions on Mobile Computing, 2024, 23(11): 10433-10459.
[33]
SongB, YuY, ZhouY, et al. Host load prediction with long short-term memory in cloud computing[J]. The Journal of Supercomputing, 2017, 74: 6554-6568.
[34]
GuptaS, DineshD A. Resource usage prediction of cloud workloads using deep bidirectional long short term memory networks[C]// Proceedings of the IEEE International Conference on Advanced Networks and Telecommunications Systems. Bhubaneswar: IEEE, 2017: 1-10.