NECO: Robust, Delay-Tolerant Sketches for Aggregating Sensor Data Streams

点击翻译按钮获取中文摘要

NeTS-NECO: Robust, Delay-Tolerant Sketches for AggregatingSensor Data StreamsEmerging data-intensive mobile sensor networking applications promisea close observation of the world around us at a relatively low dataacquisition cost. In these applications, it is a challenge toaggregate observation streams from distributed sources, due to thebandwidth and delay constraints of the network, and the energy costsof data transmission. This project develops techniques for processingvoluminous sensor data streams within the resource constraints imposedby a distributed infrastructure-less network.Techniques are designed to summarize the observation streams intosmall space "sketches". By disseminating sketches rather than the rawdata, it is possible to trade-off reduced communication cost forapproximate answers to user queries. The project investigates thetechnical challenge of making the sketches robust and suitable fortransmission over unreliable links. The sketches are delay-tolerant,i.e. they can tolerate long and variable network latencies. Variousclasses of queries are supported including: basic aggregates ofnumerical and categorical data, multi-dimensional temporal andgeographic aggregates, and time-decayed aggregates. The design ofsketches is informed and inspired by the recent rapid progress made inthe area of massive data stream processing. The project will result innew algorithms and data structures for sketching sensor data withprovable guarantees on the quality of answers. The research isexpected to benefit data-intensive sensor networking applications thatare critical to our society, such as earthquake, pollution, andtraffic monitoring.