Storage and management strategy for heterogeneous data stream based on mutation information

Mutant information For the internet of things(IoT), how to effectively store heterogeneous data streams is a new challenge. Currently random sampling is generally used for data stream storage. Additionally B+ tree is widely used to for quickly indexing. Such data in store are random, and it ignores the users' interest. In addition, B+ tree is applicable for one-dimension data, which is not feasible for multiple heterogeneous data streams. Herein, in this paper we propose a new sampling method to satisfy the users' interest according to the mutant information. Besides that an extended B+ tree structure is designed for multiple heterogeneous data stream so that the user can quickly index the interested data. Extensive experiment results show that the new sampling method and the extended B+ tree work efficiently than current sampling methods and storage mechanisms. © 2013 Springer-Verlag.