一种时空敏感的QoS预测方法
|
摘要
Web服务作为自解释的应用程序,通过在网络上提供标准接口以实现互操作性.在过去的几年中,网络上的开放服务变得越来越丰富,而服务的广泛应用导致需要高效的方法以提升性能,尤其是在软件即服务(SaaS)的场景中.目前的服务计算研究已经提出了许多基于QoS的方法,包括服务组合、服务选择、服务容错等,准确的QoS值预测显得尤为关键.服务器端的QoS值通常由服务提供商发布,代表Web服务的共享特性,而在客户端测量的QoS值则不同于服务器端:不同位置的用户可能测量到不同的QoS值.其缘由是服务调用需要消耗资源,并受到服务的状态与网络环境的影响,而且实时测量通常会引入新负载,从而导致性能评估不准确.近年来,一些方法建议引入更多的因素来提高预测精度(比如,考虑用户的感受与环境因素,引入空间及时间因素等),但是这些研究仅仅局限于某一维度的性能提升,缺少多维因素对于预测结果影响的考虑.文中提出一种时空敏感的QoS预测方法,该方法通过深度学习方法挖掘多维因素的高层特征以提高预测精度.文中在真实数据集上进行了大量实验,实验结果验证了该方法的有效性.最后,对该方法的未来发展进行展望.
Web services are self-described programmable applications conducted to achieve interoperability and accessibility over a network,and are implemented in standard interfaces and published through specific protocols.Open services on the Web become increasingly abundant in the past several years.Meanwhile,the wide-spread use of Web services requires the effective approaches,especially in Software-as-a-Service(SaaS).A lot of QoS-based approaches have been proposed for Web service composition,Web service selection,fault-tolerant Web services,etc.Accurate QoS values of Web services are desired to work well for these approaches.The QoS values of Web services can be measured both at the server-side and at the client-side.QoS values measured at the server-side are published by the service provider and represent the shared feature of Web services,which are consistent to all the users(e.g.,price,popularity,etc).However,QoS values measured at the client-side are different.Users in different locations may experience different QoS performance of Web services.Therefore,it is necessary to obtain accurate and personalized client-side Web service QoS values or their estimates.Conducting the real-world Web service evaluation at the client-side,however,is critical challenge.Web service invocations have costs.They may be charged in terms of the resources consumed in the cyberspace or the time elapse of invocations,where the server status(e.g.,workload,number of clients,etc.) and the network environment(e.g.,congestions,etc.) may change by time.Real-time performance testing may introduce overloads,which may impact the user experience of systems.Moreover,with introduced transaction workloads,the performance evaluating may not be accurate.It is difficult for various QoS-based approaches to perform well with the lack of accurate QoS values.Collaborative filtering methods are widely adopted in commercial recommender systems.QoS prediction using collaborative filtering technique was conducted firstly by Shao et al.Furthermore,advances in mobile Internet technology have enabled the clients of Web services to adjust to context changes regarding time,location and other factors.Since the services consumed by a mobile client may be different along with context changes,a multi-dimensional context model is necessary for discovering hidden relations from multi-dimensional context.Most contemporary QoS prediction methods exploit the QoS characteristics for one specific dimension,e.g.,time or location,and do not exploit multidimensional factors of context.This paper proposes a learning approach to Quality-of-Service(QoS) prediction of Web services via Spatial-Temporal context derived from the past invocation history.Our approach exploits an unsupervised encoder-decoder framework such as LSTM(Long Short Term Memory) to generate a hidden feature,which embeds the inherent characteristics for each context tracklet.Then we can calculate similarity between two context tracklets,which serves to predict Web service QoS values more accurate.To validate our approach,large-scale experiments are conducted based on a real-world Web service dataset, WSDream.The results show that our proposed approach achieves higher prediction accuracy than other approaches.Moreover,we plan to conduct more studies to predict client-side QoS properties by utilizing more other factors,and apply our approach under fog paradigm environments.The changes of application scenarios usually bring new issues.
Web services are self-described programmable applications conducted to achieve interoperability and accessibility over a network,and are implemented in standard interfaces and published through specific protocols.Open services on the Web become increasingly abundant in the past several years.Meanwhile,the wide-spread use of Web services requires the effective approaches,especially in Software-as-a-Service(SaaS).A lot of QoS-based approaches have been proposed for Web service composition,Web service selection,fault-tolerant Web services,etc.Accurate QoS values of Web services are desired to work well for these approaches.The QoS values of Web services can be measured both at the server-side and at the client-side.QoS values measured at the server-side are published by the service provider and represent the shared feature of Web services,which are consistent to all the users(e.g.,price,popularity,etc).However,QoS values measured at the client-side are different.Users in different locations may experience different QoS performance of Web services.Therefore,it is necessary to obtain accurate and personalized client-side Web service QoS values or their estimates.Conducting the real-world Web service evaluation at the client-side,however,is critical challenge.Web service invocations have costs.They may be charged in terms of the resources consumed in the cyberspace or the time elapse of invocations,where the server status(e.g.,workload,number of clients,etc.) and the network environment(e.g.,congestions,etc.) may change by time.Real-time performance testing may introduce overloads,which may impact the user experience of systems.Moreover,with introduced transaction workloads,the performance evaluating may not be accurate.It is difficult for various QoS-based approaches to perform well with the lack of accurate QoS values.Collaborative filtering methods are widely adopted in commercial recommender systems.QoS prediction using collaborative filtering technique was conducted firstly by Shao et al.Furthermore,advances in mobile Internet technology have enabled the clients of Web services to adjust to context changes regarding time,location and other factors.Since the services consumed by a mobile client may be different along with context changes,a multi-dimensional context model is necessary for discovering hidden relations from multi-dimensional context.Most contemporary QoS prediction methods exploit the QoS characteristics for one specific dimension,e.g.,time or location,and do not exploit multidimensional factors of context.This paper proposes a learning approach to Quality-of-Service(QoS) prediction of Web services via Spatial-Temporal context derived from the past invocation history.Our approach exploits an unsupervised encoder-decoder framework such as LSTM(Long Short Term Memory) to generate a hidden feature,which embeds the inherent characteristics for each context tracklet.Then we can calculate similarity between two context tracklets,which serves to predict Web service QoS values more accurate.To validate our approach,large-scale experiments are conducted based on a real-world Web service dataset, WSDream.The results show that our proposed approach achieves higher prediction accuracy than other approaches.Moreover,we plan to conduct more studies to predict client-side QoS properties by utilizing more other factors,and apply our approach under fog paradigm environments.The changes of application scenarios usually bring new issues.
引文
[1]Wolter F,Zakharyaschev M.Spatio-temporal representation and reasoning based on RCC-8//Proceedings of the 7th International Conference(KR2000).Breckenridge,USA,2000:3-14
[2]Wolter F,Zakharyaschev M.Qualitative spatio-temporal representation and reasoning:A computational perspective.Missiology:An International Review,2012,31(4):140-140
[3]Gabelaia D,Kontchakov R,Kurucz A,et al.Combining spatial and temporal logics:Expressiveness vs.complexity.Journal of Artificial Intelligence Research,2005,23(1):167-243
[4]Bennett B,Cohn A G,Wolter F,et al.Multi-dimensional modal logic as a framework for spatio-temporal reasoning.Applied Intelligence,2002,17(3):239-251
[5]Medak D I D.Lifestyles-A Paradigm for the Description of Spatiotemporal Databases[Ph.D.dissertation].Technical University Vienna,Vienna,Austria,1999
[6]Viqueira J R R.Relational algebra for spatio-temporal data management//Proceedings of the 5th International Conference on Extending Database Technology(EDBT PhD 2008).Nantes,France,2008:1-4
[7]Claramunt C.Extending Ladkin’s algebra on non-convex intervals towards an algebra on union-of regions//Proceedings of the ACM International Symposium on Advances in Geographic Information Systems.Washington,USA,2000:9-14
[8]Claramunt C,Jiang B.An integrated representation of spatial and temporal relationships between evolving regions.Journal of Geographical Systems,2001,3(4):411-428
[9]Erwig M,Schneider M.Spatio-temporal predicates.IEEETransactions on Knowledge&Data Engineering,2002,14(4):881-901
[10]Egenhofer M J.The family of conceptual neighborhood graphs for region-region relations//Proceedings of the International Conference on Geographic Information Science.Zurich,Switzerland,2010:42-55
[11]Bonarini A.The Body,the Mind or the Eye,First?.RoboCup-99:Robot Soccer World Cup III.London,UK,2000,1856:39-60
[12]Lamas-Rodrguez J,Arias J,Viqueira J R R,et al.OCURO:Estimation of Space Occupation and Vehicle Rotation in Controlled Parking Areas.Berlin,Germany:Springer,2009
[13]Zhang Xian-Feng,Cui Wei-Hong.Spatio-temporal analysis and modeling based on the integration of GIS and CA model.Journal of Image&Graphics,2000,5(12):1012-1018(in Chinese)(张显峰,崔伟宏.基于和模型的时空建模方法研究.中国图象图形学报,2000,5(12):1012-1018)
[14]Zheng Kou-Gen,Tan Shi-Yu,Pan Yun-He.A unified spatiotemporal data model based on state and change.Journal of Software,2001,12(9):1360-1365(in Chinese)(郑扣根,谭石禹,潘云鹤.基于状态和变化的统一时空数据模型.软件学报,2001,12(9):1360-1365)
[15]Deora V,Shao J,Gray W A,et al.A quality of service management framework based on user expectations//Proceedings of the 2003International Conference on ServiceOriented Computing(ICSOC).Trento,Italy,2003:104-114
[16]Zheng Z,Ma H,Lyu M R,King I.QoS-aware Web service recommendation by collaborative filtering.IEEE Transactions on Services Computing,2011,4(2):140-152
[17]Lo W,Yin J,Li Y,Wu Z.Efficient Web service QoSprediction using local neighborhood matrix factorization.Engineering Applications of Artificial Intelligence,2015,38:14-23
[18]Xiong W,Li B,He L,et al.Collaborative Web service QoSprediction on unbalanced data distribution//Proceedings of the 2014IEEE International Conference on Web Services(ICWS).Anchorage,USA,2014:377-384
[19]Zheng Z,Ma H,Lyu M R,King I.Collaborative Web service QoS prediction via neighborhood integrated matrix factorization.IEEE Transactions on Services Computing,2013,6(3):289-299
[20]Ding S,Yang S,Zhang Y,et al.Combining QoS prediction and customer satisfaction estimation to solve cloud service trustworthiness evaluation problems.Knowledge-Based Systems,2014,56:216-225
[21]Lee K,Park J,Baik J.Location-based Web service QoSprediction via preference propagation for improving cold start problem//Proceedings of the 2015IEEE International Conference on Web Services(ICWS).New York,USA,2015:177-184
[22]Chen X,Liu X,Huang Z,Sun H.Regionknn:A scalable hybrid collaborative filtering algorithm for personalized Web service recommendation//Proceedings of the 2010 IEEEInternational Conference on Web Services(ICWS).Washington,USA,2010:9-16
[23]Tang M,Jiang Y,Liu J,Liu X.Location-aware collaborative filtering for QoS-based service recommendation//Proceedings of the 2012International Conference on Web Services(ICWS).Honolulu,USA,2012:202-209
[24]Lo W,Yin J,Deng S,et al.Collaborative Web service QoSprediction with location-based regularization//Proceedings of the 2012IEEE International Conference on Web Services(ICWS).Honolulu,USA,2012:464-471
[25]Xu Y,Yin J,Lo W,Wu Z.Personalized location-aware QoSprediction for Web services using probabilistic matrix factorization//Proceedings of the 2013International Conference on Web Information Systems Engineering(WISE 2013).Hangzhou,China,2013:229-242
[26]Zhang Y,Zheng Z,Lyu M R.Real-time performance prediction for cloud components//Proceedings of the 2012IEEE International Symposium on Object/Component/Service-Oriented Real-Time Distributed Computing Workshops(ISORCW).Shenzhen,China,2012:106-111
[27]Hu Y,Peng Q,Hu X.A time-aware and data sparsity tolerant approach for Web service recommendation//Proceedings of the 2014IEEE International Conference on Web Services(ICWS).Anchorage,USA,2014:33-40
[28]Hu Y,Peng Q,Hu X,Yang R.Web service recommendation based on time series forecasting and collaborative filtering//Proceedings of the 2015 IEEE International Conference on Web Services(ICWS).New York,USA,2015:233-240
[29]Wang S,Ma Y,Cheng B,et al.Multi-dimensional QoSprediction for service recommendations.IEEE Transactions on Services Computing,2019,12(1):47-57
[30]Krizhevsky A,Sutskever I,Hinton G E.ImageNet classification with deep convolutional neural networks.Communications of the ACM,2012,60(2):2012
[31]Graves A,Mohamed A R,Hinton G.Speech recognition with deep recurrent neural networks//Proceedings of the2013IEEE International Conference on Acoustics,Speech and Signal Processing.Vancouver,Canada,2013:6645-6649
[32]Ma X,Tao Z,Wang Y,et al.Long short-term memory neural network for traffic speed prediction using remote microwave sensor data.Transportation Research Part CEmerging Technologies,2015,54:187-197
[33]Deng L,Seltzer M L,Yu D,et al.Binary coding of speech spectrograms using a deep auto-encoder//Proceedings of the Conference of the International Speech Communication Association(INTERSPEECH 2010).Chiba,Japan,2010:1692-1695
[34]Mohamed A,Dahl G E,Hinton G.Acoustic modeling using deep belief networks.IEEE Transactions on Audio Speech&Language Processing,2011,20(1):14-22
[2]Wolter F,Zakharyaschev M.Qualitative spatio-temporal representation and reasoning:A computational perspective.Missiology:An International Review,2012,31(4):140-140
[3]Gabelaia D,Kontchakov R,Kurucz A,et al.Combining spatial and temporal logics:Expressiveness vs.complexity.Journal of Artificial Intelligence Research,2005,23(1):167-243
[4]Bennett B,Cohn A G,Wolter F,et al.Multi-dimensional modal logic as a framework for spatio-temporal reasoning.Applied Intelligence,2002,17(3):239-251
[5]Medak D I D.Lifestyles-A Paradigm for the Description of Spatiotemporal Databases[Ph.D.dissertation].Technical University Vienna,Vienna,Austria,1999
[6]Viqueira J R R.Relational algebra for spatio-temporal data management//Proceedings of the 5th International Conference on Extending Database Technology(EDBT PhD 2008).Nantes,France,2008:1-4
[7]Claramunt C.Extending Ladkin’s algebra on non-convex intervals towards an algebra on union-of regions//Proceedings of the ACM International Symposium on Advances in Geographic Information Systems.Washington,USA,2000:9-14
[8]Claramunt C,Jiang B.An integrated representation of spatial and temporal relationships between evolving regions.Journal of Geographical Systems,2001,3(4):411-428
[9]Erwig M,Schneider M.Spatio-temporal predicates.IEEETransactions on Knowledge&Data Engineering,2002,14(4):881-901
[10]Egenhofer M J.The family of conceptual neighborhood graphs for region-region relations//Proceedings of the International Conference on Geographic Information Science.Zurich,Switzerland,2010:42-55
[11]Bonarini A.The Body,the Mind or the Eye,First?.RoboCup-99:Robot Soccer World Cup III.London,UK,2000,1856:39-60
[12]Lamas-Rodrguez J,Arias J,Viqueira J R R,et al.OCURO:Estimation of Space Occupation and Vehicle Rotation in Controlled Parking Areas.Berlin,Germany:Springer,2009
[13]Zhang Xian-Feng,Cui Wei-Hong.Spatio-temporal analysis and modeling based on the integration of GIS and CA model.Journal of Image&Graphics,2000,5(12):1012-1018(in Chinese)(张显峰,崔伟宏.基于和模型的时空建模方法研究.中国图象图形学报,2000,5(12):1012-1018)
[14]Zheng Kou-Gen,Tan Shi-Yu,Pan Yun-He.A unified spatiotemporal data model based on state and change.Journal of Software,2001,12(9):1360-1365(in Chinese)(郑扣根,谭石禹,潘云鹤.基于状态和变化的统一时空数据模型.软件学报,2001,12(9):1360-1365)
[15]Deora V,Shao J,Gray W A,et al.A quality of service management framework based on user expectations//Proceedings of the 2003International Conference on ServiceOriented Computing(ICSOC).Trento,Italy,2003:104-114
[16]Zheng Z,Ma H,Lyu M R,King I.QoS-aware Web service recommendation by collaborative filtering.IEEE Transactions on Services Computing,2011,4(2):140-152
[17]Lo W,Yin J,Li Y,Wu Z.Efficient Web service QoSprediction using local neighborhood matrix factorization.Engineering Applications of Artificial Intelligence,2015,38:14-23
[18]Xiong W,Li B,He L,et al.Collaborative Web service QoSprediction on unbalanced data distribution//Proceedings of the 2014IEEE International Conference on Web Services(ICWS).Anchorage,USA,2014:377-384
[19]Zheng Z,Ma H,Lyu M R,King I.Collaborative Web service QoS prediction via neighborhood integrated matrix factorization.IEEE Transactions on Services Computing,2013,6(3):289-299
[20]Ding S,Yang S,Zhang Y,et al.Combining QoS prediction and customer satisfaction estimation to solve cloud service trustworthiness evaluation problems.Knowledge-Based Systems,2014,56:216-225
[21]Lee K,Park J,Baik J.Location-based Web service QoSprediction via preference propagation for improving cold start problem//Proceedings of the 2015IEEE International Conference on Web Services(ICWS).New York,USA,2015:177-184
[22]Chen X,Liu X,Huang Z,Sun H.Regionknn:A scalable hybrid collaborative filtering algorithm for personalized Web service recommendation//Proceedings of the 2010 IEEEInternational Conference on Web Services(ICWS).Washington,USA,2010:9-16
[23]Tang M,Jiang Y,Liu J,Liu X.Location-aware collaborative filtering for QoS-based service recommendation//Proceedings of the 2012International Conference on Web Services(ICWS).Honolulu,USA,2012:202-209
[24]Lo W,Yin J,Deng S,et al.Collaborative Web service QoSprediction with location-based regularization//Proceedings of the 2012IEEE International Conference on Web Services(ICWS).Honolulu,USA,2012:464-471
[25]Xu Y,Yin J,Lo W,Wu Z.Personalized location-aware QoSprediction for Web services using probabilistic matrix factorization//Proceedings of the 2013International Conference on Web Information Systems Engineering(WISE 2013).Hangzhou,China,2013:229-242
[26]Zhang Y,Zheng Z,Lyu M R.Real-time performance prediction for cloud components//Proceedings of the 2012IEEE International Symposium on Object/Component/Service-Oriented Real-Time Distributed Computing Workshops(ISORCW).Shenzhen,China,2012:106-111
[27]Hu Y,Peng Q,Hu X.A time-aware and data sparsity tolerant approach for Web service recommendation//Proceedings of the 2014IEEE International Conference on Web Services(ICWS).Anchorage,USA,2014:33-40
[28]Hu Y,Peng Q,Hu X,Yang R.Web service recommendation based on time series forecasting and collaborative filtering//Proceedings of the 2015 IEEE International Conference on Web Services(ICWS).New York,USA,2015:233-240
[29]Wang S,Ma Y,Cheng B,et al.Multi-dimensional QoSprediction for service recommendations.IEEE Transactions on Services Computing,2019,12(1):47-57
[30]Krizhevsky A,Sutskever I,Hinton G E.ImageNet classification with deep convolutional neural networks.Communications of the ACM,2012,60(2):2012
[31]Graves A,Mohamed A R,Hinton G.Speech recognition with deep recurrent neural networks//Proceedings of the2013IEEE International Conference on Acoustics,Speech and Signal Processing.Vancouver,Canada,2013:6645-6649
[32]Ma X,Tao Z,Wang Y,et al.Long short-term memory neural network for traffic speed prediction using remote microwave sensor data.Transportation Research Part CEmerging Technologies,2015,54:187-197
[33]Deng L,Seltzer M L,Yu D,et al.Binary coding of speech spectrograms using a deep auto-encoder//Proceedings of the Conference of the International Speech Communication Association(INTERSPEECH 2010).Chiba,Japan,2010:1692-1695
[34]Mohamed A,Dahl G E,Hinton G.Acoustic modeling using deep belief networks.IEEE Transactions on Audio Speech&Language Processing,2011,20(1):14-22
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |