Abstract
We consider how an untrusted data aggregator can be assessed over multiple data streams. The aggregator could be the sink node in a sensor network where all the sensory data are gathered, or a smart-meter responsible for computing power measurements of a group of households, or any other entity that is basically in charge of answering aggregation queries such as average or summation in a data streaming environment. In these applications, important decisions are made based on the aggregated results and therefore, it is vitally important to investigate the authenticity and integrity of aggregated values. One possible approach for solving this problem is marking the data before sending it out to the aggregators (i.e. marked at the point of origin) such that the existence of those marks could be verified subsequently after the aggregation process. Our goal is to produce hidden marks that remain detectable after the aggregation and thereby not only the trustworthiness of every individual data source, but also the trustworthiness of the aggregators could be verified. This problem is referred to secure data aggregation that has been investigated by means of digital watermarking and steganography techniques in recent years. Data synchronization is a serious problem which was not addressed in the current schemes, though. Therefore, in this paper, a new watermarking construction is proposed that provides ‘synchronization marks’ in the aggregated data stream and helps protect the data itself at the end-points. Our method works at the data layer so standard transport layer security methods can be used to protect the transport of data if it is required. Finally, a set of experiments are conducted using synthesized and real sensory data as a proof of concept.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have