BigQuery exposes distributed stages, slots and shuffle. Use the execution reference below to connect those mechanics to stage work. Workload investigation applies them to query history and the 40 TB sales-table example; layout trade-offs works through BigQuery predicates and clustering.
In this chapter
- BigQuery slots, shuffle, and concurrencyStage work, join skew, reservations and runtime regressions.
For cloud-service selection, see GCP execution, messaging and storage choices.