Fault tolerance events ordering by aging learning in wireless sensor and actuator networks

Abstract

In wireless sensor and actuator networks, none of the existed events ordering algorithms addressed the network delay and failure link jointly. In ordering by double confirmation (OBDC), the process of events ordering cannot be completed, if the failure link existed in network. In temporal event ordering with fault tolerance (TEOFT), the rate of correct events ordering decreased dramatically if the network delay existed. However, both the network delay and failure link were all prone to happen over the wireless communication medium. Network delay usually varied over time in wireless network. To address this issue, the authors proposed a fault tolerance events ordering by aging learning (FTEOAL), in this study. FTEOAL defined the non-determined aging waiting time (AW) and aging learning time to order events without broadcasting any confirmation message. Thus FTEOAL could work while the network delay and failure link existed jointly. By aging learning, the AW and aging learning time were adjusted dynamically. The rate of correct events ordering (R) in FTEOAL thus could be converged to the expected R system required. The simulation results demonstrated that R in FTEOAL could be up to the expected R being 90% but R in TEOFT and OBDC was 10% and 0%.