A QoS-aware Self-adaptive System Approach In Service Computing Environment

Self-adaptive system architecture can seamlessly transfer to service computing envi ronment using service platform as an executive enabling technology. Services can enca psulate the resources, which are autonomic and independent. Services computing can s upport a flexible development of distributed applications, which includes dynamic opti mization, interoperability, evolution and customization. Self-adaptive system possesses d ecision-making mechanism and dynamic configuration architecture in order to adapt the changes of environmental context, which schedule services dynamically and provide sy stematic support for adaptive adjustment. The sense enabling technology is integrated with QoS easily. Finally, we build a QoS-aware approach based on services computing, where QoS is special because it is usually obtained by sampling method in case that we consider the cost of real-time invokings services. Thus, the efficient QoS predictio n algorithm is very necessary. Furthermore, The present study on client-oriented percei ved QoS focuses on function or performance, rarely involving energy consumption.The existing approaches have several deficiencies as follows:(1) there are few stu dies on the model which intergrates requirements and architecture, which do not consi der time changes for adaptive adjustment; (2) there are still space for QoS prediction accuracy; (3) there are few studies on QoS modeling for energy consumption. To over come the above problems, this paper focuses on the QoS-aware self-adaptive architectu re based on services computing. Our main work and contributions are as follows:(1) A self-adaptive system adapts its architecture at runtime to the changes of requirements and contexts, which assure the performance of multi-tier cloud applications deployed in virtualized server clusters according to system complexity and dynamic workloads. However, determining how to map from the requirements in the problem space to the architectural elements in the solution space is a critical research problem. This paper proposes a SAPC (self-adaptation approach based on predictive control) approach which combines requirements and architectural evolution or adaptation. It can learn a wavelet-transform-based model to predict the performance of services accurately, and induce requirements evolution or conduct architecture-based model transformations at runtime. To validate our approach, large-scale experiments are conducted based on a case study using an online SaaS platform benchmark named as CloudCRM. The results show that our proposed approach achieves higher performance than other approaches.(2) This paper proposes a collaborative approach to quality-of-service (QoS) prediction of web services on unbalanced data distribution by utilizing the past usage history of service users. It avoids expensive and time-consuming web service invocations. There existed several methods which search top-k similar users or services in predicting QoS values of Web services, but they did not consider unbalanced data distribution. Then, we improve existed methods in similar neighbors’selection by sampling importance resampling. 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.(3) This paper proposes a location-aware collaborative approach to QoS prediction of web services by utilizing the past web service usage history of service users, which avoids expensive and time-consuming web service invocations. We first acquire and process client-side spatial location information. Then, based on the collected QoS data and location information, an approach which integrates spatial location constraint and latent factor model (LFM) method considering unbalanced distribution of data is designed to achieve higher prediction accuracy for QoS value. 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.(4) There are few researches on QoS modeling for energy consumption in present, which focus on the QoS of function or performance. This paper proposes a QoS description model of energy consumption and establishes the function between resource consumption and request rate. Furthermore, this paper designed a elastic resource management approach based on perception of energy consumption in cloud computing environment. This approach accessed to information of resources cost and request changes firstly, then learned a model to predict request rate, and mapped tasks to application execution units, where resources costs are relatively fixed, in order to achieve flexible management of system energy consumption according to request rate and cost information. The results showed that our approach achieves higher performance than other approaches, which can more effectively reduce energy consumption on the premise of guaranteeing the quality of service.