Chinese Journal of Computers (计算机学报) 2013/36:12 PP.2390-2405
Since resource-limited nodes are affected by external environment factors and mobile sensor nodes exist, wireless sensor networks show dynamic characteristics. Programmer has to design redundancy mechanisms in order to eliminate the effect of dynamic characteristics. The mechanisms caused difficulty in programming sensor network. For this, we propose a novel spatiotemporal programming framework, EasiSNP. By the framework, interface reserved in each nodes can be directly programmed by EasiSNP top application which is then automatically compiled into several task blocks. Network will dynamically select suitable sensor nodes to implement task blocks through task block migration. In addition, we design EasiSNP top application translator and runtime system. The former translates top application and generates task blocks, and the latter guarantees the success of migration and execution of task block. The experiment results show that EasiSNP can build complicated application at low overhead and effectively improve robustness of network.