| Cloud computing is a new type of Internet-based computing, whereby shared resources, software, and information are provided as services to computers and other devices on demand. The architecture of the software systems involved in the delivery of cloud computing, typically involves multiple cloud components communicating with each other over application programming interfaces (API), usually implemented as Web services. Cloud computing has become a scalable service consumption and delivery platform. Web services are software systems designed to support interoperable machine-to-machine interaction over a network. The technical foundations of cloud computing include Service-Oriented Architecture (SOA), which is becoming a popular and major framework for building Web applications in the era of Web 2.0, whereby Web services offered by different providers are discovered and integrated over the Internet. Quality-of-Service (QoS) is usually employed to describe the nonfunctional properties of services in cloud and service computing. It becomes important to evaluate the QoS performance of services to differentiate the qualities of service candidates.;However, QoS evaluation is time and resource consuming. Conducting real-world evaluation is difficult in practice. Moreover, in some scenarios, QoS evaluation becomes impossible (e.g., the cloud provider may charge for service invocations, too many services to be evaluated, etc.). Therefore, it is crucial to study how to build effective and efficient approaches to predict the QoS performance of services.;In this thesis, we first propose three QoS prediction methods which utilize the users' past usage experiences. The first prediction method employs the information of neighborhoods for making QoS value prediction and engages matrix factorization techniques to enhance the prediction accuracy. The second method provides time-aware personalized QoS value prediction service. The third method employs time information for efficient online performance prediction.;The predicted QoS values can be employed to a variety of applications in cloud and service computing. We propose two applications in this thesis. The first application employs QoS information to build a Web service search engine, which help users discover appropriate Web services to fulfill both functional and non-functional requirements. The second application employs dynamic QoS information to build a robust Byzantine fault-tolerant cloud systems. |
