Time-based insertion methods for monitoring sensor data
Yang Koo Lee, Ling Wang, Young–Jin Jung, Hi-Seok Kim, Keun Ho Ryu · 2008
In wireless sensor network, the sensors collect data about natural phenomenon and transmit them to a server in real-time. Many researches mainly focus on processing continuous queries as approximate forms. However, these approaches are difficult to apply to such environmental applications which require storing the correct data. In this paper, we introduce two insertion methods, called TSI (time-segment insertion) and TPI (time-point insertion), to save the storage space without loss of raw data useful for queries by using the sensorspsila temporal attributes. In our methods, storing and discarding the incoming data is represented by a time interval that takes two timestamps at which the value of data is changed. The performance of TSI and TPI is finally judged and compared with the naive method. Because naive method does not take the temporal representation and duplicate values problem into account, the tuples obtained from data stream are quite a lot more than those of ours, whereby query execution time is greater as well.