Research On Key Technologies For Converged Services Variability

With the rapid development of networks, terminals and computer application technology, the trend of converged services development becomes providing diversification and personalization services. The converged services should support variability quite well because of individuality and diversification of user needs, and dynamism of converged network environment. Therefore, how to provide effective support for the converged service variability becomes an important issue for developing converged services.Nowadays, although service-oriented architecture is well-supported the development of converged services based on Web services, there are still many challenges in developing variable converged services, including how to manage the common and variable needs effectively, how to get the business process of services based on the user needs and how to provide services with adaptability. This attracted the attention from academia and industry. Therefore, this paper mainly in-depth research on a series of key techniques on providing variable converged services with the guidance of commonality and variability analysis. We focus on the technology and method of reuse and configuration for building services. The main contributions of this paper are summarized as follows:1. In view of the variability of converged services, combined with the idea of software product lines, the framework of Converged Service Development based on Product Lines(CSDPL) is presented, which considers the variability of developing services from the phases of requirements analysis, design and implementation. The framework establishes a bridge among changeful user needs, business processes as well as the set of Web services, explicit the main processes of developing variable converged service product lines and related concepts, and explains the relationship between the framework and the key issues of converged services which will be discussed in this thesis. The framework is the foundation for the subsequent researches.2. Time constraints are important parameters in representing a kind of variable converged services, which is difficult for scalable modeling and efficient automated checking. To overcome the above problem, this paper presents an approach for modeling and efficient verification of variable converged services. The main idea of the proposed approach is modeling behaviors of a timed service by featured timed automata, describing timed properties of service families by product Timed Computation Tree Logic(pTCTL) and analyzing the model by pTCTL modeling checking. Experiments show that our approach helps to model and verify variable converged service effectively in the early phases of development.3. For the adaptive requirements of converged services during its execution in special networks and to ensure the soundness of services after adaptation, this paper proposes a modeling and analysis method of the reliable and adaptive business process of converged services based on Markov Decision Process. The method defines the semantic similarity of business processes and the compatibility of data types, which is used to provide the candidate sets of adaptive services and makes the business process can be adaptive according to the change of circumstance. Then, the method of probabilistic model checking is been adopted to analysis the soundness and reliability of adaptive business processes. Finally, an video transfer service is been modeled and verified based on our method, which shows that our approach provides an effective underlying guideline for modeling and analyzing adaptive converged services.4. During the configuration of large-scale variable converged service, the configuration technology should take consideration of the multi-stakeholders’ quality requirements and provide an effective configuration method. In order to solve this problem, this paper proposes a variable converged service configuration method based on the fuzzy preferences measure. The comprehensive configuration criteria is been evaluated by the fuzzy comprehensive evaluation method based on the quality requirements of multi-stakeholders, which is used to construct the fitness function. Then, this paper proposes an improved binary particle swarm optimization algorithm, which is used to solve the problem of configuration. In order to configure products quickly and efficiently, an inertia weight adjustment mechanism and a velocity update strategy are introduced. Besides, a feature amends function guarantees the correctness of the solution. Experimental results show that for a large-scale variable converged service, our method outperforms traditional solutions in terms of solution time and achieves near-optimal results.