Xiangping Chen

Selecting Fault Tolerant Styles for Third-Party Components with Model Checking Support

To build highly available or reliable applications out of unreliable third-party components, some software-implemented fault-tolerant mechanisms are introduced to gracefully deal with failures in the components. In this paper, we address an important issue in the approach: how to select the most suitable fault-tolerant mechanisms for a given application in a specific context. To alleviate the difficulty in the selection, these mechanisms are abstracted as Fault-tolerant styles (FTSs) at first, which helps to achieve required high availability or reliability correctly because the complex interactions among functional parts of software and fault-tolerant mechanism are explicitly modeled. Then the required fault-tolerant capabilities are specified as fault-tolerant properties, and the satisfactions of the required properties for candidate FTSs are verified by model checking. Specifically, we take application-specific constraints into consideration during verification. The satisfied properties and constraints are evidences for the selection. A case study shows the effectiveness of the approach.

Composition of monitoring components for on-demand construction of runtime model based on model synthesis

Monitoring functionality is implemented and provided as infrastructure of most software platforms and applications. However, the construction of system model cannot be achieved by directly composing runtime model of different applications. First, models which conform to different meta-models cannot be directly composed. Second, the abstract of a system may be different from different viewpoint. Finally, the relationships between systems are different from different abstract viewpoint. In this paper, we propose a DSL (Domain Specific Language) for composition of monitoring components. The DSL distinguish component definition and instance clearly in order to enables reuse of monitoring component. Based on the component composition, the meta-model of target runtime model is generated. We developed an algorithm to generate model transformation for synthesizing runtime models of its applications. In the case study, we choose four monitoring systems as monitoring component types in our DSL. We introduce how to compose monitoring components for on-demand construction of runtime model based on model synthesis.

Using qos-contracts to drive architecture-centric self-adaptation

Self-adaptation is now a promising approach to maximize the satisfaction of requirements under changing environmental conditions. One of the key challenges for such self-adaptive systems is to automatically find a relevant architectural configuration. Existing approaches requires a set of adaptation strategies and the rough estimation of their side-effects. However, due to the lack of validation methods for such strategies and side-effects, existing approaches may lead to erroneous adaptations. Instead of side-effects, our solution leverages quality contracts whose accuracy can be separately established and which can be dynamically composed to get a quality prediction of any possible architectural configurations. To support self-adaptation, we propose a reactive planning algorithm which exploits quality contracts to dynamically discover unforeseen architectural configurations. We illustrate our approach using a running HTTP server adapting its architecture with respect to the number and the similarity of incoming requests.

An advanced operating environment for mathematics education resources

1School of Data and Computer Science, Sun Yat-Sen University, Guangzhou 510275, China; 2Academy for Intelligent Software, Guangzhou University, Guangzhou 510006, China; 3Institute of Mathematics and Computer Science, Guizhou Normal College, Guiyang 550018, China; 4School of Electronics Engineering and Computer Science, Peking University, Beijing 100871, China; 5Key Laboratory of High Confidence Software Technologies, Ministry of Education Beijing, Beijing 100871, China; 6College of Computer Sciences, New Jersey Institute of Technology, Newark 07102, USA

Towards automatic verification of web-based SOA applications

Nowadays, developing web applications in a Service-Oriented Architecture (SOA) style is emerging as a promising approach for delivering services to end users. Such web-based SOA applications are likely to suffer correctness and reliability problems mainly because their runtime environments (including web browsers and service platforms) are heterogeneous and their service interactions and flows are complex without explicit specifications. In this paper, we propose a model-checking based approach for verifying web-based SOA applications. At first, the application behavior will be automatically specified by analyzing the web-side source codes. And it will be combined with the pre-defined environment behavior so that a precise and complete enough behavior model of the application can be generated automatically. With user-defined constraint and refinement specifications, the behavior model is automatically translated to the formal specification (Promela for Spin) as the input of the model checker. If the model is flawed, the application has correctness and reliability problems. The violation traces generated by the model checker will be visualized in the behavior model for helping developers to solve the detected problems in a user-friendly manner.

A framework for the integration of MOF-compliant analysis methods

With the increasing maturity of model-driven tools and methods, new model-based analysis methods are developed to support specific stakeholder concerns during software lifecycle. This multiplication of models and their related analysis tools calls for solution addressing the integration of MOF-based analysis methods. Current research works on integration of analysis methods have already addressed the extraction of the needed input data as well as the control and the integration of the tools supporting the analysis execution. However, little attention has been paid to the integration of analysis results back into initial model. We propose a MOF-based framework enabling the integration of analysis results that a) defines a meta-model capturing the integration requirements, b) provides a MOF meta-model extension mechanism with support for upward compatibility; and c) automatically generates a model transformation for model integration. We illustrate the use of our framework by integrating a reliability analysis methods and a fault tolerant reconfiguration method on the ABC/ADL Software Architecture. We applied the resulting analysis composition onto the ECPerf JEE system.