Bachelor's thesis

My bachelor's thesis is titled "Post-Quantum key exchange protocol for IoT contexts" (orig. "Protocollo di scambio di chiavi Post-Quantistico in contesti IoT").

In this work we specialized an existing LWE-based key exchange protocol in an Internet of Things scenario, in which we assumed that the two parties had access to some additional shared approximate information which would be difficult to estimate by an adversary (for example, the communication channel quality, or some other physical quantity which depends on the physical relation between the two parties).

We show how using this assumption one can lower the sizes of the parameters used for the key exchange protocol. Note that the naive approach of using directly the shared information for the composition of the shared key is not possible as the parties do not share the exact value, rather they both have independent approximations of one unknown value.

I am fond of this work as this is realistically my first serious work in cryptography, and while the result is novel work it should be noted that it's closer to a toy-example than a groundbreaking work, as the parameter reduction is not big enough to have serious performance effects. It is, however, a good exercise in manipulation of cryptographic protocols and security proofs.

You can find the (translated) abstract of the thesis below.

Abstract

This thesis aims to specialize a key-exchange cryptographic protocol in Internet of Things (IoT) scenarios, guaranteeing security against post-quantum attacks. The security of this protocol holds via a reduction to the Ring-Learning With Errors problem. By adding the assumption of being able to utilize small physical unpredictable values detectable by the communicating parties yet hard for an adversary to estimate, such as the quality of the communication channel, we lower the sizes of the parameters involved in the protocol, albeit only slightly, making the protocol more efficient for computationally-challenged IoT devices.