Sui Academic Research Award
Sui is a Layer 1 blockchain from Sui Foundation that uses the Move language and an object centric design to execute many independent transactions in parallel, delivering high throughput and low latency. Live since May 2023, the SUI token powers gas, staking, and a storage aware fee model designed for predictable, sustainable costs. Its mission is to support the advancement and adoption of technology and Web3. In line with that mission, the Sui Foundation together with the Walrus Foundation supports academics and researchers to advance the technology and accelerate real world adoption across fields such as technology, finance, and economics.
The Sui Academic Research Awards program provides funding for research that addresses key challenges across the stack, including Systems & Security; Cryptography; Programming Languages, Developer Tooling, and Formal Verification; Economics, DeFi, and Mechanism Design; Frontier Technologies; and Walrus (data availability and storage).
Among the SARA 2025 awardees is “Lionfish: Securing Sui with Sparse Nodes,” a project by Prof. Matteo Maffei, Prof. Zeta Avarikioti, and PhD student Giulia Scaffino. As Sui’s high performance blockchain scales, operating full nodes is becoming increasingly costly in storage, computation, and bandwidth. Lionfish proposes a sparse node architecture in which nodes track and validate only targeted subsets of on chain state. By distributing validation and storage across many sparse nodes while preserving full node security guarantees, the approach lowers resource requirements, increases scalability beyond consensus and execution, and enables efficient validation, custom indexing, hotspot management, and fork detection. The result is a simpler path to node deployment and a more robust overall architecture.
Related work by Giulia Scaffino during her internship at Mysten Labs introduced Sunfish, a protocol that generalizes sparse node operation across blockchains. High throughput systems such as Sui, Aptos, and Solana can process thousands of transactions per second, which makes full nodes expensive for dApp teams, while light nodes, although cheaper, cannot ensure they observe all relevant transactions and may accept invalid ones under strong adversaries. Sparse nodes offer a middle ground: they prove completeness for all transactions that touch a selected substate and re execute those transactions to verify validity. Consequently, resource usage scales with the chosen substate rather than with the entire chain. Sunfish formalizes this model, is compatible with most existing blockchains, and reduces computational and storage costs by orders of magnitude compared with full nodes while retaining meaningful security guarantees.
Read more: Sunfish: Reading Ledgers with Sparse Nodes