Adaptive and Decentralized Operator Placement for In-Network Query Processing

摘要: In-network query processing is critical for reducing network traffic when accessing and manipulating sensor data. It requires placing a tree of operators such as filters aggregations but also correlations onto nodes in order to minimize the amount data transmitted network. In this paper, we show that problem variant task assignment which polynomial algorithms have been developed. These are however centralized cannot be used We describe an adaptive decentralized algorithm progressively refines placement by walking through neighbor nodes. Simulation results illustrate potential benefits our approach. They strategy can achieve near optimal various graph topologies despite risks local minima.