Introduction to workload management
BigQuery workload management lets you allocate and manage computeresources available for data analysis and processing, and also lets you specify howyou are charged for those resources.
Workload management models
BigQuery offers two models of workload management.Withon-demand billing, you pay for the number of bytes processedwhen you query or process your data. Withcapacity-based billing, youallocate processing capacity for workloads with the option of automaticallyscaling capacity up and down when needed.
You can switch between on-demand and capacity-based billing models at any time.You can also use acombination of the two models.
Choosing a model
Consider the following when choosing a workload management model:
| On-demand | Capacity-based | |
|---|---|---|
| Usage model | Data scanned or processed by your queries | Dedicated slots or autoscaling slots |
| Unit of metering | TiB | slot-hours |
| Minimum capacity | Up to 2,000 slots per project | 50 slots per reservation |
| Maximum capacity | Up to 2,000 slots per project | Configurable per reservation up to regional quota |
| Cost control | Optionally configure project-level or user-level quotas (hard cap) | Configure a budget expressed in slots for each reservation |
| Configuration | No configuration required | Create slot reservations and assign to projects |
| Editions support | Fixed feature set | Available in 3 editions |
| Capacity discounts | Pay-as-you-go only | Optional slot commitments for steady-state workloads |
| Predictability | Variable usage and billing | Predictable billing through baselines and commitments |
| Centralized purchasing | Per project billing | Allocate and bill slots centrally rather than for each project |
| Flexibility | Capacity on-demand (minimum 10 MiB per query) | Baseline or autoscaled slots (1 minute minimum) |
Jobs
Every time youload,export,query, orcopy data, BigQuery automatically creates, schedules, and runs a job that tracks the progress of the task.
Because jobs can potentially take a long time to complete, they run asynchronously and can bepolled for their status. Shorter actions, such as listing resources or getting metadata, are notmanaged as jobs.
For more information about jobs, seeManage jobs.
Slots
A BigQuery slot is avirtual compute unit used by BigQueryto execute SQL queries or otherjob types.During the execution of a query, BigQuery automatically determineshow many slots are used by the query. The number of slots used depends on theamount of data being processed, the complexity of the query, and the number ofslots available.
To learn more about slots and how they are used, seeunderstand slots.
Reservations
In the capacity-based pricing model, slots are allocated in poolscalledreservations. Reservations let you assign slots in ways that make sensefor your organization. For example, you might create a reservation namedprodfor production workloads, and a separate reservation namedtest for testing,so that test jobs don't compete for capacity with production workloads. Or, youmight create reservations for different departments in your organization.
For more information about reservations, seeworkload management using reservations.
BI Engine
BI Engine is a fast, in-memory analysis service that accelerates manySQL queries in BigQuery by intelligently caching the data you use mostfrequently. BI Engine can accelerate SQL queries from any source,including those written by data visualization tools, and can manage cachedtables for ongoing optimization.
BI Engine reservationsare allocated in GiB of memory and managed separately from slot reservations.
For more information about BI Engine, seeIntroduction to BI Engine.
What's next
- Understand slots
- Understand reservations
- Learn abouton-demand pricing
- Learn aboutcapacity-basedpricing
- Estimate and control costs
- Create custom cost controls
Except as otherwise noted, the content of this page is licensed under theCreative Commons Attribution 4.0 License, and code samples are licensed under theApache 2.0 License. For details, see theGoogle Developers Site Policies. Java is a registered trademark of Oracle and/or its affiliates.
Last updated 2026-02-18 UTC.