I-Corps: Stochastic Computing-based Host Intrusion Detection System

Cuny City College, New York NY

Investigators

Abstract

The broader impact/commercial potential of this I-Corps project is to provide an extra layer of security protection for Internet of Things (IoT) devices by using a novel hardware-based intrusion detection system. In today's defense-in-depth security strategy, all available security controls are placed in series aiming to protect a device from malicious cyber-physical attack. Market research estimates that the IoT market will reach over $500B by 2021 and security is considered one of the biggest barriers in adopting IoT. One main reason of this security concern is due to the limited resource of IoT devices that makes it difficult to apply all layers of security protections. The adoption of the proposed approach will provide additive security layer to the IoT devices that will boost the confidence in IoT device security in cameras, drones, sensors and others. Consequently, it may induce new industries which have previously been reluctant to apply IoT solutions to their businesses. Due to the small resource occupation of the proposed device security, even non-IoT industries may apply the proposed approach to their security applications. This I-Corps project applies Stochastic Computing (SC) to provide hardware Host Intrusion Detection System (HIDS) security solution for resource limited IoT devices. Current approaches to IoT security cannot provide HIDS solution to these resource constraint IoT devices, thereby missing out from one of the important layers in defense-in-depth security strategy. The proposed approach exploited two salient properties of SC: 1) consumption of less silicon area and power, 2) both SC and HIDS are inherently probabilistic in nature. SC is generated probabilistically from conventional number system that can perform complex mathematical operation using simple hardware blocks. The proposed approach provides IoT devices with HIDS based on simple hardware together with Machine Learning Neural Network Algorithm. Coupled with other security layers such as cryptography, access control etc., this light weight hardware solution provides strong security to the resource limited IoT devices. Embedded inside the IoT device, this HIDS monitors the incoming packets for any malicious activity and offers the last important layer of security defense. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.