Urgent.News

What's breaking now, across thousands of outlets.

Tech

Verifying formal models of Raft implementations

Antithesis demonstrated a method to verify the correctness of two Raft implementations, etcd and HashiCorp Raft, by combining AI and formal methods. They used AI, TLA+, and Antithesis to model the libraries with trace validation, iterating between model checking, Antithesis testing, and proof attempts to expose bugs in the code or gaps in the model.

They patched bugs and refined the model until it was provable and all test runs were conformant. The final proofs established core Raft safety properties, such as Election Safety, Leader Completeness, and State-Machine Safety. The process showed that specific designs in the production libraries actually offer the same guarantees as the Raft paper, even with extensions like pre-votes, snapshots, leadership transfers, read indexes, async storage writes, and membership changes. They made the artifacts available for open review.

Written by urgent.news from Lobsters's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at antithesis.com →

More in Tech

More from Friday 20 November →