Serverless Runtime / Database Co-Design with Asynchronous I/O
- Time
- 4:00 PM – 5:30 PM WAT
- Location
- Online · Discord
- Host
- Hosted by DBMSEG Project Team
We kick off our current project by reading the paper together: why embedding SQLite in a serverless runtime is so appealing, what stops it from scaling, and how asynchronous bytecode I/O and a decoupled storage engine get around it.
What we will cover
- 01Why database round trips hurt at the edge
- 02Where SQLite blocks: synchronous I/O in the bytecode engine
- 03Asynchronous bytecode instructions for I/O
- 04Decoupling the query and storage engines
- 05Reading the evaluation: where the 100× comes from
- 06Q&A, and how to join the implementation
