Multiple tasks migration in cooperative wireless network for critical services
Guangsheng Feng, Huiqiang Wang, Junyu Lin, Qian Zhao · 2013
Under complicated circumstance, random failures of wireless nodes always happen. However, some critical tasks must be running on wireless nodes to finish one job cooperatively, such as underwater data collection, forest fire sensing, and so on. Those critical tasks should running ceaselessly, and if a node failure happens, there must be a scheme to guarantee the task requirement. In response to this issue, we propose a multi-tasks migration scheme, which firstly analysis the dependency of the multiple running tasks by means of constructing a Directed Acyclic Graph (DAG), then dynamically transform those DAG tasks to interdependent and layered DAG tasks, and finally the optimal path for migration from failed nodes to healthy nodes is determined. The simulation results show that proposed scheme gives good performance to guarantee critical tasks running ceaselessly even under the situations of dynamic network workload and unknown faults.