A Dynamic Attribute-Based Load Shedding Scheme for Data Stream Management Systems

A data stream being transmitted over a network channel with capacity less than the data transmission rate of the data stream causes sequential network problems. In this paper, we present a new approach for shedding less-informative attribute data from a data stream to maintain a data transmission rate less than the network channel capacity. A scheme for shedding attributes and their data, instead of tuples, becomes imperative in data stream load shedding, since shedding a complete tuple would lead to shedding informative attribute data along with less-informative attribute data in the tuple. Our load shedding approach handles intra-stream, as well as inter- stream, load shedding such that the former sheds less-informative attribute data in a single data stream, whereas the latter sheds less-informative attribute data from multiple streams. Our load shedding approach, (i) handles wide range of data streams in different application domains, and (ii) is dynamic in nature.