Formation of adaptive subscriptions to the Internet of things data broker

Abstract

Intoduction: the possibility of interaction with the physical world through the network infrastructures of spatially distributed nodes of Internet of Things, despite the undeniable advantages of the technology, produce significant loads on information consumers. In this regard, the current interest is the creation of methods that provide the reduction of transmitted data due to the adaptive synchronization of monitoring systems with the time of real processes. One effective way to solve this problem is to use the discrete Fourier transform to determine the sampling period of the observations. Purpose: to develop an approach to the formation of adaptive data broker subscriptions based on the study of the cyclicity of observations of Internet of Things devices. Methods: the discrete Fourier transform method was applied and, based on the calculated parameters of the harmonic series, a conclusion about the frequency characteristics of the data was made. The main peaks describing the periodicity of the data are selected, the fluctuation points are determined and, according to the Kotelnikov theorem (Nyquist-Kotelnikov-Shannon Sampling Theorem), a sampling frequency that provides a sufficient intensity of observations is chosen. Results: within the corporate network of the Krasnoyarsk Scientific Center, an infrastructure of devices and applications of the Internet of Things has been deployed to monitor temperature, humidity and PM2.5 in specialized technological rooms with telecommunications equipment. The analysis showed that for different rooms the data are periodic, but their harmonic profiles do not coincide. The choice of harmonic values, the fluctuation amplitude of which determines the dynamics of changes in the observed data, should be carried out periodically for each observed device. This approach is implemented in the broker software, which distributes data in subscriptions from each of the devices in accordance with the frequency of their changes. Practical relevance: the analysis of the frequency characteristics of the data determines the broker settings, which distributes the information flows, which is one of the aspects of reliability of the IoT infrastructure.
 In addition, observing data changes will allow us to identify malfunctions in the operation of cooling systems, which can lead to the failure of complex, expensive equipment with increased heat irradiation.