The Study And Implementation On Secure Data Aggregation Of The Internet Of Things

The Internet of Things (loT) takes an important part in the forefront of communication technology, it has many important applications in military, commercial, industrial, traffic system, medical and so on. Sensor nodes forms the device layer of the Internet of Things, which has many features, such as Self-organization, large storage capacity, low power consumption and low calculation.The distribution of the nodes in device layer is intensive, a large number of wireless signal synchronous transmission will cause the network congestion and the increasing of the packet loss rate. Meanwhile, the power of nodes is limited because of the low cost and the simple structure. Too much transmission overhead will lead to rapid consumption of resources. Data aggregation technology can help decrease the transmission overhead, which is effective in improving the efficiency of the network and prolonging the life cycle of the network. But both the communication channel and the nodes are vulnerable to malicious attacks. Though data aggregation technology can decrease the transmission overhead, it also brings security threat. After analyzing the advantages and disadvantages of many security data aggregation algorithms, we suggest two new algorithms.To solve the problems of the large communication overhead of IDs in the classic algorithm CMT, we put a new secure data aggregation algorithm called EDCDSA which is based on compression coding and error correction. In this algorithm, aggregate nodes encode their child nodes’IDs by permutation and combination, and then transmit it hop by hop to the Base Station. The Base Station decodes it in a special regulation. We prove that the communication overhead in our algorithm is far less than CMT by using MATLAB.And, to solve the complex problems of distributing keys, we also put forward an approximate secure data aggregation algorithm, PADSA, in which it is unnecessary to distribute keys. With the research of the probability distribution model of the collected data, we find that adding interference data on collected data can protect the privacy of the initial data. We also prove that the Base Station can decode real data with tiny error by receiving large amount of messages, but aggregate nodes can’t decode it because the count of messages they receive is not enough.At last, we test the two algorithms in the Internet of Things Testing bed which was designed by ourselves. The result indicates that both of the two algorithms are available. And we can decrease the communication overhead on the basis of ensuring their security.