Application failure prediction method of cloud data center based on GRU Recurrent Neural Network
-
摘要: 云数据中心的分布式应用故障具有复杂性、随机性等特点，导致应用的运行与维护（简称：运维）管理任务难度大、效率低。为此，提出一种云数据中心应用故障预测方法，构建基于门控循环单元（GRU，Gated Recurrent Unit）循环神经网络（RNN，Recurrent Neural Network）的云数据中心应用故障预测模型，对云数据中心的应用监控数据进行分析处理并预测将要出现的应用故障。试验结果表明，本方法预测精确率满足应用运维管理中故障提前发现和处理的相关要求，在降低应用运维管理难度和提升运维效率方面具有一定的实用价值。
-
关键词:
- 云数据中心 /
- 循环神经网络（RNN） /
- 特征工程 /
- 门控循环单元（GRU） /
- 故障预测 /
- 单层感知器（SLP）
Abstract: Distributed application failures in cloud data centers have the characteristics of complexity, randomness, which make applications operation and maintenance tasks difficult and inefficient. Therefore, this paper proposed an application failure prediction method for cloud data center, built an application fault prediction model of a cloud data center based on GRU (Gated Recurrent Unit) Recurrent Neural Network (RNN), analyzed and processed the application monitoring data of the cloud data center and predicted the application failures that will occur. The test results showed that the prediction accuracy of this method meets the relevant requirements of failure early detection and treatment in application operation and maintenance management, and has certain practical value in reducing the difficulty of application operation and maintenance management and improving operation and maintenance efficiency. -
[1] 施卫忠. 铁路数据中心建设与规划研究 [J]. 中国铁路,2020(1):1-2. [2] 凌 政,安 藤. AIOps平台建设和实践 [J]. 网络安全和信化,2020,46(2):70-71. [3] 覃进学. 基于AIoT+AIOps的数据中心智能化运维实现探讨 [J]. 数字通信世界,2020,182(2):115. doi: 10.3969/J.ISSN.1672-7274.2020.02.083 [4] 范清栋,吴旭东,宁建创,等. 浅析人工智能在网络运维中的应用 [J]. 电脑与电信,2019(9):43-45,55. [5] 贾宇晗,李 静,贾润莹,等. 硬盘故障预测模型在大型数据中心环境下的验证 [J]. 计算机研究与发展,2015,52(S2):54-61. [6] 贾润莹,李 静,王 刚,等. 基于Adaboost和遗传算法的硬盘故障预测模型优化及选择 [J]. 计算机研究与发展,2014(S1):148-154. [7] 高建良,徐勇军,李晓维. 基于加权中值的分布式传感器网络故障检测 [J]. 软件学报,2007,18(5):1208-1217. [8] 王 成,王昌琪. 一种面向网络支付反欺诈的自动化特征工程方法 [J]. 计算机学报,2020(10):1986-1987. [9] 张超标,孙延明. 特征工程和深度前馈网络结合的刀具磨损预测 [J]. 机械设计与制造,2020,352(6):197-200. [10] 陆 航,杨涛存,刘 洋,等. 基于LSTM的动车组故障率预测模型 [J]. 中国铁路,2020(7):61-65. [11] 王明哲,张 研,杨 栋,等. 基于深度学习的车站旅客密度检测研究 [J]. 中国铁路,2019(11):13-17. [12] Yan-hua CHEN, Jian-nan LI, Kai LIAN, Qing-jie ZHU, Yi-liang LIU. Failure Prediction of Underground Pipeline Based on Artificial Neural Network[A]. Science and Engineering Research Center. Proceedings of 2017 2nd International Conference on Artificial Intelligence: Techniques and Applications (AITA 2017)[C]. Science and Engineering Research Center: Science and Engineering Research Center, 2017. [13] Greff K, Srivastava RK, Koutník J, et al. LSTM: A search space odyssey [J]. IEEE Transactions on Neural Networks and Learning Systems, 2017, 28(10): 2222-2232. doi: 10.1109/TNNLS.2016.2582924 [14] Bari M F, Boutaba R, Esteves, et al. Data center network virtualization: A survey [J]. IEEE Communications Surveys & Tutorials, 2013, 15(2): 909-928. [15] 林靖皓,秦亮曦,苏永秀,等. 基于自注意力机制的双向门控循环单元和卷积神经网络的芒果产量预测 [J]. 计算机应用,2020(S1):51-55. [16] Gaurav Singhal. LSTM versus GRU Units in RNN[EB/OL]. [2021-07-07]. https://www.pluralsight.com/guides/lstm-versus-gru-units-in-rnn.
下载:
点击查看大图
图(3)
