t-Kernel: A Translative OS Kernel for Sensor NetworksReport
The development of a reliable large-scale wireless sen- sor network (WSN) is very difficult because of resource con- straints, energy budget, and demanding application require- ments. Three OS features�OS protection, virtual memory, and preemptive scheduling�can significantly improve the reliability of WSN systems and facilitate developing com- plex WSN software. However, due to the lack of hardware support for privileged execution and address translation, it is impossible to implement these features with traditional OS design techniques. To solve this problem, we design a new OS kernel, the t-kernel, to perform extensive code modifi- cation at load time. The modified code and the OS work in a collaborative way supporting the aforementioned features. Having implemented the t-kernel on MICA2 motes, we eval- uate its performance by measuring the overhead and execu- tion speed. We analyze the CPU utilization of sensor net- work applications, and verify that, though CPU-bound tasks execute 1.5�3 times as long as in native mode, application performance under typical workloads does not noticeably de- grade. The t-kernel significantly enhances developers� abil- ity to design reliable and sophisticated sensor networks, and includes several new design techniques, such as efficient bi- nary translation on highly constrained sensor nodes, differ- entiated virtual memory without repeatedly writable swap- ping devices, and the protection of the OS from application errors without privileged execution hardware.
All rights reserved (no additional license for public reuse)
Gu, Lin, and John Stankovic. "t-Kernel: A Translative OS Kernel for Sensor Networks." University of Virginia Dept. of Computer Science Tech Report (2005).
University of Virginia, Department of Computer Science