DBOS Cloud

DBOS began as a joint research project between Stanford and MIT, advised by the creator of Postgres and the creator of Spark. In 2024, It then went on to commercialize some of the research, as well as expand upon it.

Use Cases

• Serverless Functions
• Automated Tasks
• Data Transformation
• Backend Infrastructure for Web Applications

Focus

DBOS provides a complete serverless compute platform and a set of frameworks that takes advantage of the DBOS Cloud. Using the frameworks you can write durable functions for the cloud, with guaranteed once and only once execution, workflows, time-based execution, and observability built right in. DBOS is truly serverless, charging only for usage and scaling to zero, not even charging for CPU wait time.

Benchmarks

DBOS published a benchmark showing their serverless architecture executes durable workflows 25x faster than AWS Lambda & Step Functions. They accomplish this impressive feat using a lightweight library instead of an external orchestrator. You can also run a benchmark yourself with the instructions in their GitHub repository.