Cloud-assisted spatio-textual k nearest neighbor joins in sensor networks

k nearest neighbors (kNN) query is an important problem in a variety of sensor network applications. Traditionally, we handle this problem with a single query processing approach, which just considers the location information. It usually neglects the other information such as temperature, humidity, pressure, etc. In order to overcome the defect of the traditional approaches, we investigate the problem from a new perspective and desire to solve a more interesting problem called spatio-textual k nearest neighbor join (ST-kNNJ). It searches text-similar and k-nearest sensors to a query set containing more than one query point. With the help of cloud computing, ST-kNNJ can be processed in distributed computational environment to gain better processing capability and response efficiency. In this paper, we generalize the problem of ST-kNNJ and propose our approaches to it. And we can deal with large-scale data when using MapReduce framework. Evaluation results show that our approach achieve better performance in comparison with the naive approach.