Algorithm for identifying clients based on dynamic MAC addresses in narrowly targeted secure networks using deep learning neural networks

Abstract

Existing algorithms for identifying clients on a network segment are based on static client binding by MAC address. MAC address generation is based on pseudo-random sequences of 0–256 characters. With this feature in mind, software was developed based on an algorithm for generating MAC addresses using bidirectional neural networks, followed by integration of a decision support system module. A secondary feature of the developed software is the ability to set a MAC validity timeout, which will limit access to the network segment and increase the security factor.