Computing over encrypted spatial data generated by IoT

Proliferation of IoT devices produces the enormous amount of data that need to be stored on clouds. A main focus of this paper is to ensure the secrecy of data, while it is in transit through unsecure communication media from edge devices to cloud and to provide security to the data once it is stored on public cloud. A new techniques based on computing over encrypted data i.e. homomorphic encryption seems to be promising methods. However, most of the previous works supporting computing over encrypted data are neither efficient nor compatible with IoT data and real time spatial data streams because there is a huge difference between normal data and spatial data. In this paper, we proposed a framework for computing over IoT generated encrypted spatial data. In order to provide computation over encrypted data first it needs to be indexed in standard data structure. For indexing encrypted data, we used R tree and its variants. We also proposed a method of most efficient and scalable, parallel construction of R trees and its variants on real time encrypted spatial data. We fired spatial range queries on encrypted spatial data. Specifically, the spatial range query execution time over encrypted spatial data of our proposed scheme is extremely efficient which takes slightly more time as taken by normal spatial range query executed over non-encrypted real time spatial data. Our scheme is not only efficient, but also highly compatible and scalable with IoT generated spatial data. Moreover, we rigorously define the scalability, query performance time, analyze the security of our schemes, and also conduct extensive experiments with a real time spatial dataset to demonstrate the performance of our schemes.