This analysis examines a 2008 research paper that proposed building a database directly on Amazon S3, separating storage from compute decades before it became a cloud-native standard. The paper's design used SQS as a write-ahead log and S3 for page storage, employing asynchronous checkpointing to overcome S3's high latency. While the implementation details are dated, the core philosophy of disaggregated architecture strongly parallels modern systems like Amazon Aurora.
Background
Traditional databases historically coupled storage and compute resources, but cloud-native architectures have increasingly adopted disaggregated designs. Amazon S3 (Simple Storage Service) is AWS's object storage service, while SQS (Simple Queue Service) is a message queuing service.
- Source
- Lobsters
- Published
- Mar 6, 2026 at 01:45 AM
- Score
- 6.0 / 10