Talks

PhD Proposal: Data Structures and Protocols to Improve Scalability and Security of Distributed Consensus

Shravan Srinivasan

5165 Brendan Iribe Center for Computer Science and Engineering (IRB)

Thursday, January 27, 2022, 1:00-3:00 pm

You are subscribed to this talk through .
You are watching this talk through .
You are subscribed to this talk. (unsubscribe, watch)
You are watching this talk. (unwatch, subscribe)
You are not subscribed to this talk. (watch, subscribe)

Abstract

Cryptographically secure append-only decentralized ledgers, blockchains, are envisaged as a platform for various applications ranging from financial transactions to digitally bound contracts. However, despite achieving internet-scale consensus in the permissionless setting, current blockchain protocols face several scalability and security issues. This thesis proposes new authenticated data structures and distributed consensus protocols to improve scalability (through stateless blockchains) and security (through support for more powerful network adversaries) of blockchains, respectively.

In the first part, I present Hyperproofs, a Vector Commitment (VC) scheme to improve scalability through stateless blockchains. Since VCs provide a short commitment to a set of values, they can remove the storage required to validate transactions and realize a stateless design. Our VC is the first construction that is efficiently maintainable (can update all proofs in sublinear time) and aggregatable (can combine multiple individual proofs into a single succinct proof). Hyperproofs also incentivize proof computation through a new property called unstealability, which allows a prover to cryptographically bind the proofs she computes with her identity irreversibly. As a subsequent work, I aim to study bilinear cryptographic accumulators that can succinctly prove the membership and non-membership of multiple elements efficiently in the distributed setting for privacy-preserving applications.

In the second part, I discuss my proposed work on designing secure consensus protocols against powerful network adversaries that can delay or delete any message in the network. Specifically, I aim to: (1) design a permissionless protocol in a weaker synchronous model called the mobile sluggish model, where a subset of the honest nodes can arbitrarily lose synchrony, and (2) design a round efficient Byzantine broadcast protocol under strongly adaptive and majority corruptions.

Examining Committee:

Chair:
Department Representative:
Members:

Prof. Charalampos (Babis) Papamanthou
Prof. John Dickerson
Prof. Jonathan Katz

Bio

Shravan Srinivasan is a Ph.D. student in the Department of Computer Science at the University of Maryland, advised by Prof. Papamanthou. His research focuses on applied cryptography and distributed systems.

This talk is organized by Tom Hurst