分布式用户隐私保护可调节的云服务个性化QoS预测模型

doi:10.11959/j.issn.2096-109x.2023022

网络与信息安全学报 ›› 2023, Vol. 9 ›› Issue (2): 70-80.doi: 10.11959/j.issn.2096-109x.2023022

分布式用户隐私保护可调节的云服务个性化QoS预测模型

许建龙, 林健, 黎宇森, 熊智

汕头大学工学院，广东汕头 515063

修回日期:2023-02-21 出版日期:2023-04-25 发布日期:2023-04-01
作者简介:许建龙（1982- ），男，广东汕头人，汕头大学讲师，主要研究方向为隐私计算、服务计算、区块链
林健（1996- ），男，福建莆田人，汕头大学硕士生，主要研究方向为隐私计算、服务计算
黎宇森（2000- ），男，广东佛山人，主要研究方向为隐私计算、服务计算
熊智（1978- ），男，湖北黄冈人，汕头大学教授，主要研究方向为信息安全、分布式计算
基金资助:
国家自然科学基金(61702318);广东省基础与应用基础研究基金(2023A1515010707);广东省基础与应用基础研究基金(2021A1515012527);广东省科技专项资金（“大专项+任务清单”）项目(STKJ2021201);广东省普通高校重点领域专项(2022ZDZX1008)

Distributed user privacy preserving adjustable personalized QoS prediction model for cloud services

Jianlong XU, Jian LIN, Yusen LI, Zhi XIONG

School of Engineering, Shantou University, Shantou 515063, China

Revised:2023-02-21 Online:2023-04-25 Published:2023-04-01
Supported by:
TheNational Natural Science Foundation of China(61702318);Guangdong Province Basic and Applied Basic Research Fund(2023A1515010707);Guangdong Province Basic and Applied Basic Research Fund(2021A1515012527);Guangdong Science and Technology Special Fund Project （“Major Project + Task List”）(STKJ2021201);Special Projects in Key Fields of Guangdong Universities(2022ZDZX1008)

摘要/Abstract

摘要：

个性化服务质量（QoS，quality of service）预测是构建高质量云服务系统的重要环节，传统基于协同过滤方法采用集中式的训练模式难以保护用户隐私，为了在获取高准确预测效果的同时有效保护用户隐私，提出分布式用户隐私保护可调节的云服务个性化QoS预测模型（DUPPA）。该模型采用“服务器-多用户”架构，服务器协调多个用户，处理多用户上传模型梯度和下载全局模型的请求并维护全局模型参数。为进一步保护用户隐私，提出用户隐私程度调节策略，通过调节本地模型参数初始化比例、梯度上传比例以平衡隐私程度和预测精度。在本地模型初始化阶段，用户计算本地模型与全局模型的差值矩阵，并选择差值矩阵中数值较大元素所对应的全局模型参数初始化本地模型参数；在梯度上传阶段，用户可选择部分重要的梯度上传至服务器来满足不同应用场景对隐私保护的需求。为了评估 DUPPA 的隐私程度，提出针对分布式矩阵分解模型梯度共享方案的数据重构攻击方法。实验结果表明，当DUPPA在梯度上传比例为0.1、本地模型参数初始化比例为0.5时，预测的平均绝对误差（MAE，mean absolute error）和均方根误差（RMSE， root mean square error）较传统的集中式矩阵分解模型分别降低了1.20%和0.91%；当DUPPA的梯度上传比例为0.1时，隐私程度至少是梯度上传比例为1时的5倍；当DUPPA的本地模型参数初始化比例为0.5时，隐私程度至少是本地模型参数初始化比例为1时的3.44倍。

关键词: 云服务, 隐私保护, 分布式矩阵分解, 服务质量预测

Abstract:

Personalized quality of service （QoS） prediction is crucial for developing high-quality cloud service system.However, the traditional collaborative filtering method, based on centralized training, presents challenges in protecting user privacy.In order to effectively protect user privacy while obtaining highly accurate prediction effect, a distributed user privacy adjustable personalized QoS prediction model for cloud services （DUPPA） was proposed.The model adopted a “server-multi-user” architecture, in which the server coordinated multiple users, handled multiple users’ requests for uploading model gradients and downloading global model, and maintained global model parameters.To further protect user privacy, a user privacy adjustment strategy was proposed to balance privacy and prediction accuracy by adjusting the initialization proportion of local model parameters and gradient upload proportion.In the local model initialization stage, the user calculated the difference matrix between the local model and the global model, and selected the global model parameters corresponding to the larger elements in the difference matrix to initialize the local model parameters.In the gradient upload stage, the user can select some important gradients to upload to the server to meet the privacy protection requirements of different application scenarios.To evaluate the privacy degree of DUPPA, a data reconstruction attack method was proposed for the distributed matrix factorization model gradient sharing scheme.The experimental results show that when DUPPA sets the gradient upload proportion to 0.1 and the local model parameter initialization proportion to 0.5, the predicted MAE and RMSE are reduced by 1.27% and 0.91%, respectively, compared with the traditional centralized matrix factorization model.Besides, when DUPPA sets the gradient upload proportion to 0.1, the privacy degree is 5 times higher than when the gradient upload proportion is 1.And when DUPPA sets the local model parameter initialization proportion to 0.5, the privacy degree is 3.44 times higher than when the local model parameter initialization proportion is 1.

Key words: cloud service, privacy protection, distributed matrix factorization, quality of service prediction

中图分类号:

TP309

许建龙, 林健, 黎宇森, 熊智. 分布式用户隐私保护可调节的云服务个性化QoS预测模型[J]. 网络与信息安全学报, 2023, 9(2): 70-80.

Jianlong XU, Jian LIN, Yusen LI, Zhi XIONG. Distributed user privacy preserving adjustable personalized QoS prediction model for cloud services[J]. Chinese Journal of Network and Information Security, 2023, 9(2): 70-80.

图/表 8

图1

表1

表2

表3

表4

图2

图3

表5

参考文献 31

[1]	ZHU J , HE P , ZHENG Z ,et al. A privacy-preserving qos prediction framework for web service recommendation[C]// Proceedings of 2015 IEEE International Conference on Web Services. 2015: 241-248.
[2]	CHEN Z , SUN Y , YOU D ,et al. An accurate and efficient web service QoS prediction model with wide-range awareness[J]. Future Generation Computer Systems, 2020,109: 275-292.
[3]	ZHANG Y , PAN J , QI L ,et al. Privacy-preserving quality prediction for edge-based IoT services[J]. Future Generation Computer Systems, 2021,114: 336-348.
[4]	LIU J , CHEN Y . A personalized clustering-based and reliable trust-aware QoS prediction approach for cloud service recommendation in cloud manufacturing[J]. Knowledge-Based Systems, 2019,174: 43-56.
[5]	WU D , HE Q , LUO X ,et al. A posterior-neighborhood-regularized latent factor model for highly accurate web service QoS prediction[J]. IEEE Transactions on Services Computing, 2019.
[6]	YANG Y , ZHENG Z , NIU X ,et al. A location-based factorization machine model for Web service QoS prediction[J]. IEEE Transactions on Services Computing, 2018,14(5): 1264-1277.
[7]	LI J , WANG J , SUN Q ,et al. Temporal influences-aware collaborative filtering for QoS-based service recommendation[C]// Proceedings of 2017 IEEE International Conference on Services Computing (SCC). 2017: 471-474.
[8]	WU H , YUE K , LI B ,et al. Collaborative QoS prediction with context-sensitive matrix factorization[J]. Future Generation Computer Systems, 2018,82: 669-678.
[9]	CHEN C , LIU Z , ZHAO P ,et al. Privacy preserving point-of-interest recommendation using decentralized matrix factorization[C]// Proceedings of the AAAI Conference on Artificial Intelligence. 2018,32(1).
[10]	牛犇, 李凤华, 华佳烽 ,等. 移动网络中场景关联的隐私保护机制研究[J]. 网络与信息安全学报, 2015,1(1): 31-42.
	NIU B , LI F H , HUA J F ,et al. Research on scenario-based mechanism in privacy-aware mobile networks[J]. Chinese Journal of Network and Information Security. 2015,1(1): 31-42.
[11]	VOIGT P , VON DEM BUSSCHE A . The eu general data protection regulation （gdpr）[M]. A Practical Guide （1st Ed）. Cham: Springer International Publishing, 2017(3152676):10.5555.
[12]	SHAO L , ZHANG J , WEI Y ,et al. Personalized QoS prediction forweb services via collaborative filtering[C]// Proceedings of IEEE International Conference on Web Services (ICWS 2007). 2007: 439-446.
[13]	SARWAR B , KARYPIS G , KONSTAN J ,et al. Item-based collaborative filtering recommendation algorithms[C]// Proceedings of the 10th international conference on World Wide Web. 2001: 285-295.
[14]	ZHENG Z , MA H , LYU M R ,et al. QoS-aware web service recommendation by collaborative filtering[J]. IEEE Transactions on services computing, 2010,4(2): 140-152.
[15]	ZHENG Z , MA H , LYU M R ,et al. Collaborative web service QoS prediction via neighborhood integrated matrix factorization[J]. IEEE Transactions on Services Computing, 2012,6(3): 289-299.
[16]	ZHANG Y , ZHANG X , ZHANG P ,et al. Credible and online qos prediction for services in unreliable cloud environment[C]// Proceedings of 2020 IEEE International Conference on Services Computing (SCC). 2020: 272-279.
[17]	ZHU X , JING X Y , WU D ,et al. Similarity-maintaining privacy preservation and location-aware low-rank matrix factorization for QoS prediction based web service recommendation[J]. IEEE Transactions on Services Computing, 2018.
[18]	张鹏程, 金惠颖 . 一种移动边缘环境下面向隐私保护 QoS 预测方法[J]. 计算机学报, 2020,43(8): 1555-1571.
	ZHANG P C , JIN H Y . A forward privacy-preserving QoS prediction method for mobile edge environments[J]. Chinese Journal of Computers. 2020,43(8): 1555-1571.
[19]	BADSHA S , YI X , KHALIL I ,et al. Privacy preserving location-aware personalized web service recommendations[J]. IEEE Transactions on Services Computing, 2018.
[20]	BADSHA S , YI X , KHALIL I ,et al. Privacy preserving user based web service recommendations[J]. IEEE Access, 2018,6: 56647-56657.
[21]	RAHMAN M S , KHALIL I , ALABDULATIF A ,et al. Privacy preserving service selection using fully homomorphic encryption scheme on untrusted cloud service platform[J]. Knowledge-Based Systems, 2019,180: 104-115.
[22]	ZHANG S , LIU Q , LIN Y . Anonymizing popularity in online social networks with full utility[J]. Future Generation Computer Systems, 2017,72: 227-238.
[23]	QI L , ZHANG X , LI S ,et al. Spatial-temporal data-driven service recommendation with privacy-preservation[J]. Information Sciences, 2020,515: 91-102.
[24]	ZHANG Y , ZHANG P , LUO Y ,et al. Efficient and privacy-preserving federated QoS prediction for cloud services[C]// Proceedings of 2020 IEEE International Conference on Web Services (ICWS). 2020: 549-553.
[25]	ZHU L , LIU Z , HAN S . Deep leakage from gradients[J]. Advances in Neural Information Processing Systems, 2019,(32).
[26]	RUDER S . An overview of gradient descent optimization algorithms[J]. arXiv preprint arXiv:1609.04747, 2016.
[27]	LIU D C. , NOCEDAL J . On the limited memory BFGS method for large scale optimization[J]. Mathematical Programming, 1989,45(1): 503-528.
[28]	ZHENG Z , ZHANG Y , LYU M R . Investigating QoS of real-world web services[J]. IEEE Transactions on Services Computing, 2012,7(1): 32-39.
[29]	WEI W , LIU L , LOPER M ,et al. A framework for evaluating gradient leakage attacks in federated learning[J]. arXiv preprint arXiv:2004.10397, 2020.
[30]	MNIH A , SALAKHUTDINOV R R . Probabilistic matrix factorization[J]. Advances in Neural Information Processing Systems, 2007,20.
[31]	MCMAHAN B , MPPRE E , RAMAGE D ,et al. Communication-efficient learning of deep networks from decentralized data[C]// Proceedings of Artificial Intelligence and Statistics. 2017: 1273-1282.

上传比例	MD = 2.5%		MD = 5.0%		MD = 7.5%		MD = 10.0%
上传比例	MAE/RMSE	MSE	MAE/RMSE	MSE	MAE/RMSE	MSE	MAE/RMSE	MSE
1	0.634/1.609	0.000 1	0.550/1.40	0.000 1	0.512/1.301	0.000 1	0.500/1.290	0.002
0.1	0.650/1.633	0.000 5	0.577/1.422	0.001	0.520/1.296	0.002	0.504/1.278	0.124
0.01	0.694/1.723	0.090	0.615/1.477	0.712	0.531/1.302	2.806	0.532/1.280	2.067
0.001	0.835/1.948	6.402	0.715/1.621	9.556	0.650/1.406	14.052	0.650/1.396	14.185

重写比例	MD = 2.5%		MD = 5.0%		MD = 7.5%		MD = 10.0%
重写比例	MAE/RMSE	MSE	MAE/RMSE	MSE	MAE/RMSE	MSE	MAE/RMSE	MSE
1	0.650/1.633	0.000 5	0.577/1.422	0.001	0.520/1.296	0.002	0.504/1.278	0.124
0.5	0.652/1.642	0.043	0.573/1.414	0.193	0.528/1.322	0.095	0.500/1.265	0.426
0.1	0.654/1.660	1.847	0.588/1.454	3.757	0.536/1.332	1.271	0.515/1.281	1.660
0.05	0.672/1.679	4.393	0.630/1.549	7.454	0.594/1.432	5.146	0.586/1.335	2.615

上传比例	θ _o=1		θ _o=0.5		θ _o=0.1		θ_o=0.05
上传比例	MAE/RMSE	MSE	MAE/RMSE	MSE	MAE/RMSE	MSE	MAE/RMSE	MSE
1	0.500/1.290	0.002	0.496/1.275	0.071	0.507/1.278	0.682	0.564/1.320	2.761
0.1	0.504/1.278	0.124	0.500/1.265	0.426	0.515/1.281	1.660	0.586/1.335	2.615
0.01	0.532/1.280	2.067	0.539/1.267	2.810	0.556/1.304	9.929	0.583/1.346	8.294
0.001	0.650/1.396	14.185	0.654/1.373	13.553	0.647/1.376	16.118	0.689/1.490	27.348

模型	MD = 2.5%		MD = 5.0%		MD = 7.5%		MD = 10.0%
模型	MAE	RMSE	MAE	RMSE	MAE	RMSE	MAE	RMSE
CMF	0.652 3	1.650 2	0.580 4	1.427 5	0.528 3	1.321 3	0.500 9	1.268 2
DUPPA	0.652 9	1.642 7	0.573 0	1.414 5	0.528 0	1.322 5	0.500 8	1.265 2
增益	-0.09%	0.45%	1.27%	0.91%	0.05%	–0.09%	0.01%	0.23%

用户比例	MD = 2.5%	MD = 5.0%	MD = 7.5%	MD = 10.0%
1	56	42	35	28
0.5	112	86	71	54
0.2	354	226	164	139
0.1	568	548	398	282

分布式用户隐私保护可调节的云服务个性化QoS预测模型

Distributed user privacy preserving adjustable personalized QoS prediction model for cloud services

在线阅读

pdf下载

可视化

摘要/Abstract

引用本文

使用本文

图/表 8

参考文献 31

相关文章 15

Metrics

推荐阅读 0

[1]	陈赛特, 李卫海, 姚远志, 俞能海. 轻量级K匿名增量近邻查询位置隐私保护算法[J]. 网络与信息安全学报, 2023, 9(3): 60-72.
[2]	肖敏, 毛发英, 黄永洪, 曹云飞. 基于属性签名的车载网匿名信任管理方案[J]. 网络与信息安全学报, 2023, 9(2): 33-45.
[3]	孙哲, 宁洪, 殷丽华, 方滨兴. 基于教学实训靶场的“数据隐私保护”课程建设初探[J]. 网络与信息安全学报, 2023, 9(1): 178-188.
[4]	白雪, 秦宝东, 郭瑞, 郑东. 基于SM2的两方协作盲签名协议[J]. 网络与信息安全学报, 2022, 8(6): 39-51.
[5]	肖敏, 姚涛, 刘媛妮, 黄永洪. 具有隐私保护的动态高效车载云管理方案[J]. 网络与信息安全学报, 2022, 8(6): 70-83.
[6]	卢晨昕, 陈兵, 丁宁, 陈立全, 吴戈. 具有紧凑标签的基于身份匿名云审计方案[J]. 网络与信息安全学报, 2022, 8(6): 156-168.
[7]	明盛智, 朱建明, 隋智源, 张娴. 信息增值机制下在线医疗隐私保护策略[J]. 网络与信息安全学报, 2022, 8(6): 169-177.
[8]	张娴, 朱建明, 隋智源, 明盛智. 数字货币交易匿名性与监管的博弈分析[J]. 网络与信息安全学报, 2022, 8(5): 150-157.
[9]	刘峰, 杨杰, 齐佳音. 区块链密码学隐私保护技术综述[J]. 网络与信息安全学报, 2022, 8(4): 29-44.
[10]	金琳, 田有亮. 基于区块链的多权限属性隐藏电子病历共享方案[J]. 网络与信息安全学报, 2022, 8(4): 66-76.
[11]	张伟成, 卫红权, 刘树新, 普黎明. 5G移动边缘计算场景下的快速切换认证方案[J]. 网络与信息安全学报, 2022, 8(3): 154-168.
[12]	陈前昕, 毕仁万, 林劼, 金彪, 熊金波. 支持多数不规则用户的隐私保护联邦学习框架[J]. 网络与信息安全学报, 2022, 8(1): 139-150.
[13]	高振升, 曹利峰, 杜学绘. 基于区块链的访问控制技术研究进展[J]. 网络与信息安全学报, 2021, 7(6): 68-87.
[14]	杨冠群, 刘荫, 徐浩, 邢宏伟, 张建辉, 李恩堂. 基于区块链的电网可信分布式身份认证系统[J]. 网络与信息安全学报, 2021, 7(6): 88-98.
[15]	宋甫元, 秦拯, 张吉昕, 刘羽. 基于访问控制安全高效的多用户外包图像检索方案[J]. 网络与信息安全学报, 2021, 7(5): 29-39.