Authors:

Kovid T

Date Created:

Not specified

Course Title:

Professor:

Not specified

About Paper:

Topology-Hiding Computation (THC) aims to develop secure multi party computation (MPC) protocols for incomplete communication networks, while hiding the underlying network topology. THC protocols have various applications, including anonymous networks and sensor placement. This work considers THC under a fail-stop adversary, in which the adversary can crash any party at any point of time. In prior works, proposed protocols for THC under fail-stop adversaries only guarantee security with abort, which gives an adversary the ability to control which nodes receive their output, while also leaking information about the network's topology. Furthermore, proposed protocols fail to extend to the asynchronous setting, since they crucially rely on detecting crashes. This research aims to further understand protocols for THC tolerating fail-stop corruptions in both the synchronous and asynchronous setting with minimal leakage. In the synchronous setting, we provide a protocol that tolerates fail-stop adversaries with Guaranteed Output Delivery (GOD) for all graphs. In the asynchronous setting with polynomial median delays, we provide a protocol that tolerates up to t fail-stop corruptions with no leakage for t+1 connected graphs. We also provide a novel protocol in the asynchronous setting for a wider range of graph classes and delay distributions based on proxy broadcasters. Our protocols are based on a combination of novel cryptographic primitives such as secure hardware & threshold encryption. This is a joint work with Aniket Kate, Chen-Da Liu-Zhang, and Satvinder Singh,. Keywords: Applied Cryptography; Security & Privacy; Anonymity

Source:

Purdue University / 2025

Topics:

No topics listed

Co-authors:

Kovid T