PhD Defense: Data Structures and Protocols for Scalability and Security of Distributed Consensus
Shravan Srinivasan
Tuesday, May 16, 2023, 2:00-4:00 pm 
Abstract
Distributed consensus is the problem of reaching an agreement among mutually distrusting nodes in the presence of faults. Despite tremendous progress in achieving internet-scale consensus, prior consensus protocols face scalability and security challenges. This dissertation proposes new authenticated data structures and protocols to improve the scalability (through stateless
In the first part, I will present our Vector Commitment (VC) data structure, Hyperproofs, to improve the scalability of blockchains. 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 succinct proof). Hyperproofs also incentivize proof computation through a new property called unstealability, which allows a prover to cryptographically bind the proofs she computes irreversibly with her identity. I will also present schemes to succinctly prove and verify the (non-)membership of multiple elements in a cryptographic accumulator for applications in the distributed setting.
In the second part, I will show how we improve the resilience of consensus protocols against powerful network adversaries that can delay or delete any message in the network. Specifically, I will describe the first permissionless protocol that is secure even after relaxing the standard synchrony assumption. Finally, I will present the first constant-round Byzantine broadcast under a strongly adaptive and corrupt majority setting.
In the first part, I will present our Vector Commitment (VC) data structure, Hyperproofs, to improve the scalability of blockchains. 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 succinct proof). Hyperproofs also incentivize proof computation through a new property called unstealability, which allows a prover to cryptographically bind the proofs she computes irreversibly with her identity. I will also present schemes to succinctly prove and verify the (non-)membership of multiple elements in a cryptographic accumulator for applications in the distributed setting.
In the second part, I will show how we improve the resilience of consensus protocols against powerful network adversaries that can delay or delete any message in the network. Specifically, I will describe the first permissionless protocol that is secure even after relaxing the standard synchrony assumption. Finally, I will present the first constant-round Byzantine broadcast under a strongly adaptive and corrupt majority setting.
Examining Committee
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