2026-05-09 –, Sladkovský Stage
An analysis of the first 100 days of the Ethereum Fusaka hard fork, based on real-world PeerDAS mainnet data. We examine PeerDAS performance and data availability, discuss partial messages, EIP-8077, and the sharded blob mempool, and explain the future of PeerDAS and why 2D encoding is needed.
This talk presents an analysis of the first 100 days of the Ethereum Fusaka hard fork, with a focus on real-world PeerDAS behavior and data observed on mainnet. We will examine how PeerDAS has performed in practice, including data availability, network overhead, and validator participation, and highlight key insights derived from empirical measurements. Building on this analysis, the talk explores the future evolution of PeerDAS, explaining why 2D encoding is a necessary next step to improve scalability, robustness, and fault tolerance. We will also cover partial messages in the context of PeerDAS, their role in data propagation and recovery, and how they interact with upcoming protocol changes such as EIP-8077. Finally, we discuss the design and implications of the sharded blob mempool, and how these components together shape the roadmap for scalable data availability on Ethereum.
Dr. Leonardo Bautista-Gomez is the founder and research leader of MigaLabs. He is also a Core Researcher for Ethereum. He has been collaborating with the Ethereum Foundation for over seven years, receiving multiple research grants from the EF, as well as other institutions. He has over a decade of research experience in supercomputers, deep learning, and blockchain technology.