Using a Deep-Learning Approach for Smart IoT Network Packet Analysis

The Internet of Things is an emerging technology of network devices. Most IoT devices are enabled to be connected to external cloud servers purposely or incidentally, and our cell phones can also connect to the same cloud to make and apprentice such IoT devices. It is highly likely that an attack goes into or comes from IoT devices. Since IoT devices appear or disappear with no notification or no authentication, typical firewall rule-based intrusion detections are unsatisfiable. This paper describes an artificial neural network approach to training a network packet inspector to identify potentially attackable packets to/from IoT devices. The contribution of this paper includes 1) the characterization of IoT attacks, and 2) the formation of neural network nodes for each attack. This paper also shows the simulation results of neural network performance evaluation.

Paper

Full text

PDF

Using a Deep-Learning Approach for Smart IoT Network Packet Analysis

Semantic Scholar · Computer Science · 2019

Abstract

The Internet of Things is an emerging technology of network devices. Most IoT devices are enabled to be connected to external cloud servers purposely or incidentally, and our cell phones can also connect to the same cloud to make and apprentice such IoT devices. It is highly likely that an attack goes into or comes from IoT devices. Since IoT devices appear or disappear with no notification or no authentication, typical firewall rule-based intrusion detections are unsatisfiable. This paper describes an artificial neural network approach to training a network packet inspector to identify potentially attackable packets to/from IoT devices. The contribution of this paper includes 1) the characterization of IoT attacks, and 2) the formation of neural network nodes for each attack. This paper also shows the simulation results of neural network performance evaluation.

Similar papers

© 2026 NYSGPT2525 LLC