Abstract: Theoretical results in cryptography and distributed computing are always proved relative to some formal model that is intended to accurately capture the real world. We focus here on the effect of different network models on the feasibility of various tasks related to consensus. First, we highlight assumptions about the power of adaptive adversaries in distributed protocols, noting that while in one model adaptively secure broadcast is impossible when more than half the parties may be corrupted, in another model adaptive security is possible for any number of corruptions. Next, we discuss the synchronous and asynchronous communication models for network communication and argue that neither provides a perfect model for real-world traffic. We then introduce a new “hybrid” protocol for consensus that is resilient to some fraction of corruptions when the network is synchronous and remains secure (though for a lower threshold of corruptions) even when the network is asynchronous. Our protocol is optimal in the number of corruptions it tolerates.
Bio: Jonathan Katz is a professor of computer science at the University of Maryland, and director of the Maryland Cybersecurity Center. He is a co-author of the widely used textbook “Introduction to Modern Cryptography,” now in its second edition, as well as a monograph on digital signature schemes. He has served as the program chair for the annual Crypto conference as well as the ACM Conference on Computer and Communications Security, and as a member of the steering committee for the IEEE Cybersecurity Initiative and the State of Maryland Cybersecurity Council. He currently serves as an editor for the Journal of Cryptology.
Katz is a recipient of a Humboldt Research Award as well as the University of Maryland Distinguished Scholar-Teacher award. He was recently named an International Association for Cryptologic Research (IACR) Fellow.