Adaptive Holistic Scheduling for In-Network Sensor Query Processing

We observe two problems in the current scheduling schemes for in-network sensor query processing: (1) A query execution plan never changes after it is injected into the network and (2) the data communication schedule rarely considers the query workload. Both problems severely hurt the performance, because the runtime dynamics, such as the wireless connectivity and the data flows, change frequently and affect the performance greatly. To address these two problems, we propose an adaptive, holistic scheduler, AHS, which schedules both the query operators and the data communication, and is able to adapt the schedules to the runtime dynamics. We have implemented AHS in nesC and have tested it on real motes as well as in simulation. Our results show that AHS improves the performance of query processing in various dynamic settings.