Study On Key Technologies For Web Service Discovery

The Internet not only changes our ways of study, work, and living, but also enriches our spiritual world. However, complexity inevitably arises with the convenience brought by the Internet. In the application level of Web service, increasing number of service types and continuous proliferation of services pose a number of problems on the healthy development of Web service ecosystem. Nowadays, service providers have been deploying thousands of homogenous services targeting various similar application scenarios. Hence, developing scalable and effective means for an automated service discovery with minimal human intervention in various application environments is a great challenge of Web service technology.This dissertation conducts a research on the key technologies for Web service discovery. Based on the analysis of service quality model, service discovery manner, and service discovery architecture, the prevalent service discovery techniques are thoroughly reviewed. For various network environments and application scenarios, the advantages and disadvantages of different service discovery models are elaborated. A series of innovative solutions are brought forward to approach existing problems. The major contributions are outlined as follows:We present a traffic management model for Web service discovery in distributed environment. The proposed model analyzes the node behavior through the operation of five queues. The traffic of a node is regulated by a traffic management module. Different traffic policies are applied to the query queue, answer queue, and forward queue. The priorities of the above three queues were adjusted for the purpose of improving the availability and latency of service discovery. The experimental data demonstrated that the forward queue is most important; hence it should be assigned with the highest priority. The second most important queue is the answer queue and the least important queue is the query queue.In the area of non-functional attributes based service selection, various users have different priorities of Qo S attributes. We propose a collaborative Qo S-aware service evaluation method which is based on the Analytical Hierarchy Process theory(AHP). The user preference is mapped into the hierarchical structure of the AHP, and is reflected by the weights of different criteria. The Qo S data is given by past users according to their usage experiences. Since the malicious users and biased users are inevitable, we introduce a trust threshold to provide reputation management. Hence, the Qo S data is collected from trusted users. In order to further guarantee the reliability of the collected Qo S data, the outliers are excluded with statistical analysis. The service with the highest benefit to cost ratio among candidate services is selected through the aggregation of the collected Qo S data.We present a k-median facility location agent for low-cost service selection in digital community network. It is conventional that users in a digital community network take part in a collaborative mechanism for the purpose of service discovery and service selection. This inevitably brings unnecessary burdens to users. We invent a new way which considers providing an overall service selection result for users in the community from the perspective of the administrators of a digital communication network. The proposed model analyzes the service selection process in a digital communication network through five entities and six types of messages. In order to reduce the total connection cost between service requestors and the facilities where services are deployed, we develop a local search algorithm and a greedy algorithm. When a compromise between the total connection cost of the digital communication network and the number of facilities connected to the digital communication network is needed, there exists an inflection point for the number of facilities connected to the digital communication network.We develop a self-governing privacy-preserving dynamic service discovery architecture for mobile ad hoc networks. The architecture is able to adjust its parameters based on the statuses of the mobile ad hoc network and autonomically switch the working mode between directory-based and directory-less. We introduce a network model and a unified service information management scheme. In directory-based mode, the service registration/deregistration is introduced. The key generation and the searching mechanism for service query are established. For the purpose of privacy preservation for a node and topology control, we develop a direction-probing algorithm, a local location optimization algorithm, and a global location optimization algorithm. A function tuning algorithm is designed to approach the problem of energy conservation for a node. In addition, this algorithm helps to prolong the lifetime of the service discovery network. In directory-less mode, a two-hop zone scheme is introduced to analyze the service discovery process. Then, the level of connectivity of a node is described. The level of connectivity for a network is modeled subsequently. The status of a mobile ad hoc network and the related parameters are monitored by a monitoring token. With regard to three fundamental performance indicators of service discovery: availability, message overhead, and delay, we conduct extensive experiments to evaluate the proposed model. The experimental data demonstrated that the autonomic mode switch function could significantly improve the performance of service discovery.