Devising a method for stabilizing control over a load on a cluster gateway in the internet of things edge layer

The object of this study is the process of managing overload at the boundary layer of the geographically distributed Internet of Things. The task addressed is reducing the number of losses of information packets of the geographically distributed Internet of Things arriving at the boundary layer gateway. For this purpose, it was proposed to use fast temporal horizontal scaling and stabilizing control over the load formed in the gateway buffer. In the process of conducting research, a temporal horizontal scaling algorithm was developed for the boundary layer cluster gateway. The evolutionary Firefly Algorithm with a fitness function based on the Lorenz function was used in the development. This made it possible to speed up the search for a cluster node for operational temporal scaling of the gateway for the period of gateway buffer overload. The standard algorithm for intelligent queue management has been modified. The modification is based on the proposed method for stabilizing load control over the boundary layer cluster gateway of the geographically distributed Internet of Things. The method takes into account the features of the boundary layer architecture. The proposed method made it possible, in case of reaching the upper threshold of the queue, to scale the gateway until its buffer is filled. As a result, the number of information packet losses arriving at the boundary layer gateway was reduced. Studies of the proposed method showed that the number of information packet losses is reduced compared to existing methods. The research results can be explained by the use of temporary horizontal scaling of the gateway and fixing the thresholds of the information packet queue buffer. The method is effective at an average load level on the cluster gateway from 0.2 to 1.2