Grant Application – Zebra State Snapshot & Fast Sync Infrastructure

State Snapshot and Fast Sync infrastructure

Decision text was not parsed from the minutes.

The project will build a state snapshot and fast-sync module for Zebra, the Zcash Rust full node, enabling new nodes to bootstrap from verified point-in-time state snapshots instead of replaying the entire blockchain from genesis. This infrastructure component will export, compress, verify, distribute, and import finalized chain state without changing consensus rules or protocol behavior. By dramatically reducing node startup time and resource costs, it will improve operator experience, enable efficient CI/CD and testing workflows, and support downstream analytics systems. The module will provide secure, multi-level integrity verification, a custom segmented snapshot format, resume-capable downloads, and CLI tools, delivered as a reusable, modular crate integrated with zebra-state. Applicant is requesting $30,000. Gguy: I understand that there is a problem with creating fast syncing infrastructure which is beneficial in specific scenarios, mainly in testing and development, and that’s important. It’s my understanding that there are specific solutions that different teams use to achieve this but encouraging this usage more widely as this application proposes is probably not something ZCG should fund. Without greater protections, likely utilizing zero knowledge proofs, it’s hard to imagine that lessening the burden of syncing by having snapshots is good for the security of the network. So with those concerns I’m going to vote to reject this application. Artkor: While the idea itself is attractive, from my perspective this approach increases centralization, making it unlikely to be included in Zebra’s automatic new-node path and therefore unlikely to see broad adoption. For this reason, I reject this proposal. And I found the alternative approach outlined by Hahn during the discussion more compelling. Hanh: Well if you really want to have new infrastructure for development on demand like in Infura, they should build something different because this is not going to save much time. This is still going to take hours to boot. But the real question is do you really need to have an infrastructure like Infura since Zcash does not have smart contracts. So people don’t develop in the same way in Zcash. For most of the development, a reg test is enough and that can be started in a matter of minutes. If they want to speed up the installation of a new node they can copy from another node that they have and if they don’t have that then they should take it from a full blockchain and just wait. My position is the risk inferred by skipping a ton of checks or implementing your own checks outweighs the advantage of saving a little bit of time. Zerodartz: Yes, saving time to start a full node would be nice. But for example with a slower connection speed the difference would be rather minor. This is a reject from me now, but hopefully in the future Zcash nodes can cut down in some way the sync time. decentralistdan: Agreed, reject. Unanimous rejection