Security of Nakamoto Consensus under Congestion
The security of Nakamoto’s consensus protocol, which powers blockchains like Bitcoin and Ouroboros, relies on nodes being able to download and process blocks in time. Previous works have proven Nakamoto consensus secure assuming a synchronous network model in which all blocks are processed within a bounded delay. However, this model neglects the queuing delays resulting from bursty block production and delayed release of blocks by the adversary. In this talk, I will show an attack, called the “teasing attack”, which exploits queuing delays to overcome previously proven security bounds. I propose a more realistic bounded-bandwidth network model in which nodes can process a certain number of blocks per second. I will present new bounds on the security of Nakamoto consensus, derived using new probabilistic and combinatorial analysis methods in the bounded-bandwidth model.
This talk is based on the following two papers: https://arxiv.org/abs/2111.12332 and https://arxiv.org/abs/2303.09113
Online Link (for remote participation): http://tinyurl.com/bdhcdpp4
Password: bwRBRu
Speaker Biography
Srivatsan Sridhar is currently a 5th-year PhD candidate at Stanford University, where he is advised by Professor David Tse. His research primarily focuses on the security and scalability of consensus protocols. He has published at top-tier security and blockchain conferences such as NDSS, IEEE S&P, AFT, and Financial Crypto. His industry experience includes working as a security researcher at Protocol Labs, where he addressed key vulnerabilities in IPFS, the largest peer-to-peer filesystem, and at Subspace Labs where he was instrumental in the development of a new proof-of-storage blockchain. Prior to joining Stanford, Srivatsan did his B.Tech. in electrical engineering at IIT Bombay, where he undertook research on secure multiparty computation and signal processing, and received the President of India gold medal and the Prof. K. C. Mukherjee award for the best B.Tech. project in 2019.
