High availability is a key requirement in mobile communication systems, especially, when it is used for mission-critical services such as public safety. A failure in the fixed network infrastructure can affect a large number of its users and risk loss of lives. The software applications that run the infrastructure and provide services to the mobile users according to some defined communication protocols play a key role for system availability. However, the real-time peer-to-peer aspects of these protocols pose a real challenge to the development of a recovery method that can work in such environment.

In a previous paper, we introduced a new failure recovery approach based on a behavioral model for communication protocols. We extend our model based recovery method to also include infrastructure applications that provide data delivery with QoS. Our experimental results have shown that the method improves the availability of these applications by enabling them to quickly and successfully recover from failures. Moreover, the failure-free overhead caused by this approach is relatively low, and is experimentally found to be less than 5%.