Corelab Seminar
2020-2021
Vassilis Zikas
Proof-of-Reputation Blockchain with Nakamoto Fallback
Abstract.
Reputation is a major component of trustworthy systems.
However, the subjective nature of reputation, makes it tricky to base a
system's security on it. In this work, we describe how to leverage
reputation to establish a highly scalable and efficient blockchain. Our
treatment puts emphasis on reputation fairness as a key feature of
reputation-based protocols. We devise a definition of reputation
fairness that ensures fair participation while giving chances to newly
joining parties to participate and potentially build reputation. We also
describe a concrete lottery in the random oracle model which achieves
this definition of fairness. Our treatment of reputation-fairness can be
of independent interest. To avoid potential safety and/or liveness
concerns stemming from the subjective and volatile nature of reputation,
we propose a hybrid design that uses a Nakamoto-style ledger as a
fallback. To our knowledge, our proposal is the first cryptographically
secure design of a proof-of-reputation-based (in short PoR-based)
blockchain that fortifies its PoR-based security by optimized
Nakamoto-style consensus. This results in a ledger protocol which is
provably secure if the reputation system is accurate, and preserves its
basic safety properties even if it is not, as long as the fallback
blockchain does not fail.
This is joint work with Leonard Kleinrock and Rafail Ostrovsky.
Bio: Vassilis Zikas is joining Purdue University to lead its
blockchain-lab initiative. Prior to joining Purdue, Vassilis was an
Associate Professor and Vice-Director of the Blockchain Technology
Laboratory at the University of Edinburgh, and a research fellow and
area leader for multi-party computation at R&D blockchain company IOHK.
He has also held academic faculty, visiting, and senior researcher
positions at RPI, Simons Institute for the Theory of Computing, UC
Berkeley, UCLA, and ETH Zurich, and was a Research Fellow of the Simons
Institute, UC Berkeley.