Kirincas2: An Alternative Software Implementation of a Data Aggregator in a Sensor Network and its Effects on Performance

This thesis describes an implementation of a data aggregator and Internet gateway for a waterpipe monitoring system. It aggregates sensing data from a Controller Area Network (CAN) of sensing nodes distributed along the surface of a water pipe, and it performs data logging with a time stamp to a local flash storage device. Current implementations of the data aggregator have been limited in scalability by the relatively low data logging performance. We overcome this limitation by organizing the software into tasks and interrupt service routines with the proper scheduling by a lightweight real-time operating system (RTOS). Experimental results on a complete implementation of this data aggregator show that it is indeed able to achieve higher performance without data loss among implementations on low-power hardware of the same class.