eNOMT promises to make Substrate‑based blockchains more scalable and cost‑effective by shrinking proof sizes and cutting SSD I/O, enabling higher transaction throughput on commodity hardware. This could broaden validator participation and accelerate the adoption of parachain ecosystems.
Gabriele introduced eNOMT, a Nearly Optimized Merkle Trie designed to replace Substrate’s existing 16‑ary Merkle trie with a binary variant that dramatically reduces proof size and read‑write amplification. By encapsulating merkelization within a flat key‑value store (B3) and leveraging binary branching, eNOMT cuts sibling nodes from 45 to 12 for comparable depths, enabling single‑SSD‑access reads and near‑native SSD latency for state queries.
The talk detailed three core advantages: smaller proofs that free up space in Proof‑of‑Validity (PoV) payloads, easier scaling on consumer‑grade SSDs without terabyte‑scale RAM, and faster chain performance through reduced I/O. Technical hurdles—such as Substrate’s variable‑length keys and long shared prefixes—were addressed via a forked eNOMT implementation, new B3 encoding, collision‑key handling, transparent hashing, and “jump pages” that collapse long prefix paths into single disk pages.
A proof‑of‑concept built on a Polkadot fork demonstrated functional reads, writes, forks, and RPC queries using eNOMT, achieving a working prototype in two weeks. Next steps include benchmarking, adapting proof formats for the relay chain, and integrating the new trie into Cumulus for parachains. If successful, eNOMT could lower hardware barriers for validators, increase transaction throughput per block, and streamline decentralized network operation.
Comments
Want to join the conversation?
Loading comments...