Spacetime Addresses Scalability, Focusing on Horizontal Scaling for OLTP Workloads

Understanding Scalability in Spacetime

Spacetime detailed its strategy for scalability, a common concern for database systems. The company distinguishes between vertical and horizontal scalability, emphasizing that horizontal scalability, which involves adding more computers to handle increased load, is often what users expect from a scalable system.

Dimensions of Scaling

The company identifies three core dimensions of scaling: compute, storage, and networking. While horizontally scaling storage is considered relatively straightforward and is slated for release by October 31, 2026, Spacetime notes that horizontal scaling for all aspects of networking and computation is not universally applicable or efficient.

Challenges with OLTP Databases

Spacetime specifically addresses the complexities of scaling OLTP databases. It highlights that general horizontal scalability claims in OLTP databases can lead to significant overhead per transaction and poor performance when transactions contend for resources. Spacetime aims to provide high performance even under contention and offers tools to facilitate scaling of parallelizable OLTP workloads.

Comparison to Other Systems

The discussion frequently references CockroachDB as a representative example of general-purpose, horizontally scaling relational database management systems, including technologies like Spanner and Aurora DSQL. Spacetime acknowledges the engineering achievements of these systems while outlining its own approach to address specific scaling challenges, particularly in high-contention OLTP scenarios.