← All stories
● Covered by 1 source · 1 reportMedium impact1 neutral

Cursor Develops Continuity, a Git Storage Architecture Using S3-Backed WAL for Scalability

🔄 Updated 1d ago
New to BrevFeed? We gather this story from every outlet covering it into one summary — ranked by real-world impact, not just the latest headline — so you never miss what matters. What is BrevFeed? →

Key points

  • Continuity uses S3-backed WAL as the source of truth for Git storage.
  • Achieves over 300 pushes per second using S3 Express One Zone.
  • Supports linear read scaling with up to 100 replicas in tests.
  • Local NVMe repositories act as warm caches, not authoritative copies.

Continuity Architecture Overview

Cursor has introduced Continuity, a new Git storage architecture designed to scale Git operations. Unlike traditional systems that rely on replica coordination for consistency, Continuity uses an S3-backed write-ahead log (WAL) as the primary source of truth. This fundamental shift aims to improve scalability and performance for Git hosting.

Performance and Scalability

Continuity demonstrates linear read scaling, supporting up to 100 replicas in synthetic tests. Cursor reports achieving more than 300 pushes per second when utilizing S3 Express One Zone. This performance targets both large repositories with intensive continuous integration (CI) workloads and the numerous smaller repositories generated by coding agents.

Technical Implementation

In the Continuity model, pushed data is stored in S3, and corresponding reference updates are recorded in the WAL. A push is acknowledged only after the necessary data has been durably persisted. Local NVMe repositories function as warm caches, with the S3-backed WAL remaining the authoritative copy. UDP gossip propagates WAL updates, and conditional S3 reads verify replica state, with these reads averaging under 10 milliseconds.

Comparison to Existing Solutions

Existing Git hosting solutions, such as GitHub's Spokes architecture, use multiple NVMe replicas and a three-phase commit for reference updates. While this keeps replicas synchronized, coordination overhead increases with more replicas. Continuity's approach of treating Git storage like a database, with S3 as the durable source, offers a different model for handling consistency and scalability.

Industry Reception

The Continuity architecture has drawn comparisons to database systems due to its focus on push consistency, force-push transactions, and point-in-time recovery. Casey Lee, CTO at Liatrio, noted the architectural shift of using S3 as the source of truth and local NVMe as a cache, though he also mentioned that Cursor's performance figures have not been independently verified.

✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →

The daily brief

One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.

One email a day. Unsubscribe in one click, any time.

Today's brief

Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.

~34 min · 27 stories · Oct 02

▶ Play today's brief Listen on Spotify

New every morning, and the back catalogue is archived by date.

Reporting from

Cursor developed Continuity, a Git storage architecture that uses an S3-backed write-ahead log (WAL) as the source of truth, moving away from replica coordination for consistency. This design allows for linear read scaling and achieves over 300 pushes per second, addressing the needs of large repositories and those used by coding agents.