Study On Online Detection And Resolution Method For Web Services Feature Interaction

Feature interaction problem, firstly coined in the telecom by Bellcore, is one of the important obstacles for the development of new services. But this phenomenon is not unique to the telecom. It also occurs in some software system beyond telecom. In the Web services area, service composition technique is introduced to integrate several Web services to meet the increasing user demand that a single service can not satisfy. But it also brings the potential of the problem resembling the feature interaction in the telecommunication area. The problem is defined as Web Services Feature Interactions (WSFI) problem. Although this problem was raised in May 2000, it has not received enough attention. The related work is limited, and there exists some deficiencies. For example, the application fields of current detection methods are limited and they can not apply to the online scenarios.According to the research trends in this area, the thesis is focused on the online detection and resolution of WSFI problem, and makes an in-depth study on how to manage the problem in a uniform and online manner by taking inspirations from the perfect characteristics of immune system and related immune principles.The main contributions of our work are discussed in this thesis in detail, and they are summarized as follows:1. An immune-inspired online detection system for WSFI problem (WSFIDS) is presented.Aiming to overcome the drawbacks of current research work in the Web Services Feature Interactions area, an immune-inspired online detection system for WSFI problem (WSFIDS) is presented by drawing inspirations from the immune theory. It employs the multi-layer defense mechanism, and overcomes the limitation of classification-based approach so as to detect the WSFI phenomena in a uniform and online manner. With the application of some immune principles including antigen presentation, antigen recognition, co-stimulation, immune memory and so on, the system can effectively detect not only the known WSFI, but also the unknown ones.2. A formal description framework for WSFI is proposed.Formal description of WSFI phenomena conduces to the further study on the essence of WSFI problem, and helps grasp what is essential for WSFI detection leading to propose an effective detection method. A formal description framework for WSFI is proposed by case study in this thesis, and used as one of the important bases in the WSFIDS system.3. A clone selection-based optimization algorithm for WSFI detection rules is presented.WSFI detection rules play an indispensable role in the implementation of WSFI detection method. Inspired by the clone selection theory, a clone selection-based optimization algorithm for WSFI detection rules is presented according to the experimental results of our detection method. It can effectively decrease the detection time so as to increase the detection efficiency.4. An online resolution method for WSFI is proposed.The aim of WSFI detection is resolution. By analyzing the related mechanism of immune system, the immunity-based WSFI resolution mechanism is briefly discussed in this thesis. And a transaction rollback-based WSFI resolution method is proposed to resolve the specific type of WSFI resulting from the atomic service invalidity at runtime.