SAFE^d: Self-Attestation For Networks of Heterogeneous Embedded Devices

09/18/2019 ∙ by Alessandro Visintin, et al. ∙ 0

The Internet of Things (IoT) is an emerging paradigm that allows a fine-grained and autonomous control of the surrounding environment. This is achieved through a large number of devices that collaboratively perform complex tasks. To date, IoT networks are used in a variety of critical scenarios and therefore their security has become a primary concern. A robust technique to enhance the integrity of remote devices is called Remote Attestation (RA). However, classic RA schemes require a central and powerful entity, called Verifier, that manages the entire process of attestation. This makes the entire system dependent on an external entity and inevitably introduces a single point of failure. In our work, we present SAFE^d: the first concrete solution to self-attest autonomous networks of heterogeneous embedded devices. SAFE d overcomes the limitations of the previous works by spreading the duties of the Verifier among all the devices in a scalable way. In our schema, the information needed for the attestation phase is replicated inside the network, thus raising the bar to accomplish an attack. As a result, the IoT network can self-inspect its integrity and self-recover in case of attack, without the need of an external entity. Our proposal exploits the security guarantees offered by ARM TrustZone chips to perform a decentralized attestation protocol based on an enhanced version of a distributed hash table. We implemented a prototype of SAFE d for the Raspberry Pi platform to evaluate the feasibility and the security properties of our protocol. Moreover, we measured the scalable properties of SAFE d by using a large network of virtual devices. The results show that SAFE d can detect infected devices and recover up to 99.7 Moreover, we managed to protect 50 devices with a logarithmic overhead on the network and the devices' memory.



There are no comments yet.


page 1

This week in AI

Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.