Cursor Uses S3 WAL to Scale Git Storage to More than 300 Pushes per Second
Cursor has introduced Continuity, a Git storage architecture that uses an S3 backed write ahead log as the source of truth. The design turns local NVMe repositories into warm caches and separates replica coordination from consistency. Cursor reports linear read scaling with up to 100 replicas and more than 300 pushes per second with S3 Express One Zone in synthetic tests. By Leela Kumili
Cursor has developed Continuity, a Git storage architecture that utilizes an S3-backed write-ahead log (WAL) as the source of truth, instead of relying on replica coordination for consistency. This new architecture enables linear read scaling with up to 100 replicas in synthetic tests and performance of over 300 pushes per second using S3 Express One Zone.
The Continuity model changes the traditional replication and compaction scaling approach used in Git hosting. Instead of coordinating multiple replicas, it makes the S3-backed WAL the durable source of truth, allowing local repositories to converge independently on the state stored in S3. This change shifts the consistency boundary from the replica layer to durable object storage, trading replica coordination for storage validation, asynchronous propagation, and extra bandwidth.
Cursor reports that reads take less than 10 milliseconds on average, and the system remains consistent even if the gossip communication fails. The storage model has drawn comparisons with database systems, as it treats Git storage like a database, with considerations for push consistency, force-push transactions, and point-in-time recovery.
Written by urgent.news from InfoQ's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
