Architecture | SurrealDB

In this section, we will explore the SurrealDB Cloud architecture and how it is designed to provide a scalable, high-performance, and secure database solution.

SurrealDB Cloud benefits from SurrealDB's layered architecture, which separates storage from compute, enabling improved scalability, durability and availability without the need to operate, manage, scale or shard your database.

Start

For development and staging applications with vertically scalable requirements. SurrealDB Cloud Start provides dedicated storage and a single compute node that you can resize as workloads grow.

Scale

For business-critical applications that must stay available when a node fails and scale query throughput horizontally. Scale clusters run on SurrealDS, SurrealDB's distributed storage engine — with quorum consensus, compute–storage separation, and object-storage-backed durability — and start at three compute units. The storage tier handles replication and distributed transactions; query nodes can be added or replaced without copying the full dataset to each machine.

Provisioning a production cluster on Cloud avoids the operational work of self-hosting: Kubernetes, replication, patching, backups, and upgrade orchestration are managed for you.

Note

Some Scale capabilities — such as read replicas, database branching, and database forking — are still rolling out. See Pricing and the Scale waitlist for current availability.

To learn more about deployment models and self-hosted SurrealDS, see Deployment and SurrealDS.