Mithril

Mithril is a Solana full node client written in Golang with the goal of serving as a "verifying full node" with lower hardware requirements than that of Solana validators and RPC nodes. This project is being developed upon the foundations of Radiance, which was built by Richard Patel (@ripatel) with contributions from @leoluk.

This project is under active development. Our aim is to move rapidly with Mithril's development, hence all code is likely to be incomplete, buggy, and/or improperly tested at any particular point in time. Please check the dev branch for the latest version of the codebase. We'll begin to add documentation as Mithril begins to approach a useable state.

Milestone 1 (Completed): Reimplementation of the Solana Virtual Machine in Golang

• Completed in August 2024, read more here.
• Reimplementation of all syscalls, with a comprehensive test suite developed and exercised; bugs found as a result fixed.
• Reimplementation of all native programs, with a comprehensive test suite developed and exercised; bugs found as a result fixed.
• Implementation of the remainder of the runtime and VM, with a comprehensive test suite also developed. Any bugs found as a result of testing and review to be fixed.

Milestone 2 (In progress): Block Replay and Simple RPC Interface

• Snaphot retrieval and loading (Completed).
• Full implementation of transaction (and therefore block) handling.
• Incorporate the AccountsDB and Blockstore facilitie that are necessary for data storage and retrieval.
• Work on minimal RPC interface.
• Development and intensive use of a robust and comprehensive ‘conformance suite’ for verification of compliance of the VM, interpreter, and runtime as a complete unit. Differential fuzzing will be used to detect differences versus relevant versions of the Labs client, and guided fuzzing will be used generally to uncover security and loss-of-availability issues. Any bugs identified during this phase will be remediated.

Milestone 3 (Future): Intensive Work on System Optimization

• Thorough optimization work on entire system, including on components such as the Virtual Machine and AccountsDB.
• Implementation of block batch processing with configurable block window size.

Potential Future Directions

• Broaden block retrieval sources beyond gossip and RPC node services (TBD how this is implemented).
• Implementation of ‘archival node’ features which would include building out historical replay compatibility.