2025-06-13 –, Side Stage - Cinema 7
A new design that mixes a sharded mempool with data availability sampling to achieve higher rollup throughput and high scalability for Ethereum. We present the results of our research in the P2P network for efficient blob dissemination and bandwidth reduction.
Together with EF researchers we have worked on a new design for sharding the blob space, improved the way blobs are disseminated over the network, and designed a robust data sampling strategy that can withstand large scale byzantine attacks. We will present changes at the P2P layer that are necessary to optimize the way blob data is disseminated. New techniques to reduce redundancies between the execution layer and the consensus layer while ensuring high security. The end game of this research is to demonstrate that it is possible to dramatically increase the blob space, hence the throughput of Ethereum without decreasing the decentralization level of the network.
Dr. Leonardo Bautista-Gomez is the founder and research leader of MigaLabs and a Core researcher of Ethereum. He has been collaborating with the EF for over seven years and has received multiple research grants from the EF, and other institutions. He has over a decade of research experience in supercomputers, deep learning and blockchain technology. He has published more than 50 scientific articles and has received multiple international academic awards such as the IEEE TCSC Award for Excellence in Scalable Computing and the ACM/IEEE George Michael Memorial High Performance Computing Fellow. He did his Ph.D. at the Tokyo Institute of Technology and his Master's at the Pierre & Marie Curie Paris 6 University.