On The Outside Looking In: Towards Detecting Counterfeit Devices Using Network Traffic Analysis

Abstract

In order to combat the growing threat of counterfeit components, this paper describes a technique to fingerprint and identify components based upon the way in which they send and process ICMP, UDP, and TCP network packets. These network fingerprints are then processed by an artificial neural network in order to categorize and classify individual processors. These tests found that ICMP and UDP packets provided an an effective, inexpensive, and fast tool for identifying counterfeit components in a network.