DuckDB has an answer to “bloated” database messages

Older database protocols were designed before the era of big data and often transmitted results row by row, including unnecessary details like column names and types for each row. This approach, according to Guillermo Sánchez Dionis, product team lead at DuckLabs, results in "quite a bloated message" particularly when handling large query results.

DuckDB aims to address this issue with Quack, an HTTP-based protocol previewed in May and released in October as part of DuckDB 2.0.0 LTS. Quack allows developers to start working with results even while a query is still executing, a feature Sánchez Dionis emphasizes as a significant advantage. Unlike older protocols that serialize data one value at a time, Quack leverages DuckDB's columnar representation and vectorized execution, sending results as a stream.

This design minimizes memory footprint and enables concurrent writing from multiple clients, thanks to DuckDB's multi-version concurrency control. Quack's efficiency gains are expected to continue with upcoming DuckDB 2.0 patch releases.