A two-level strategy for software fault tolerance of service control point

Fault tolerance (FT) is a feature of top importance in the long-life real-time systems. The service control point (SCP) in the intelligent network (IN) is just one of this type of systems. Its dependability affects directly the quality of service (QoS) of the whole network. We give an overview on the software fault tolerance (SFT) at first, and decide to adopt the time-redundancy approach to implement the SFT of the SCP according to the special features of the IN application. Based on the processing mechanism of IN calls and with the consideration of the load-balance, software fault-tolerance and software upgrading on-line, we present a two-level concurrency model of the SCP software, and give our implementation of this model in the multi-task environment. Furthermore, we discuss in detail the issues of the SFT based on this model, and give the corresponding strategies and algorithms to achieve the specified whole FT targets. Finally, we discuss the system effectiveness of the SCP software, and provide a method of computing the effectiveness of the SCP software based on the Markov model