Access via BigQuery

Customers can query their Zus data mart directly via Google BigQuery. The data remains in Zus's AWS account and BigQuery runs federated queries against the Zus tables, such that the experience is the same as querying tables hosted in BigQuery.

What this feature enables

• Query your Zus data in BigQuery. Same tables and columns as the Zus Snowflake share, so a query you already have mostly carries over.
• Join it to your own data. You can join Zus tables to your own tables in BigQuery. See the cost note below before you do this a lot.
• Connect a BI tool. Anything that talks to BigQuery - Looker, Tableau, Sigma, Hex - connects the normal way.
• Give access to your own team. Use your own Google Groups to manage access in your organization.

Limits of this feature

• It's read-only. No writing, no creating tables in the Zus dataset, no scheduled jobs inside our project.

What you need before you start

1. A Google Cloud project with billing turned on. Your queries run in your project and get billed to you, so it needs a working billing account.
2. A supported BigQuery edition. You need either plain on-demand billing or the Enterprise edition. Standard and Enterprise Plus cannot query this data at all. That's a Google restriction on the cross-cloud feature, not a Zus setting. You can find your edition in the BigQuery console under Capacity management → Reservations, or ask whoever owns your Google Cloud billing.
3. Google accounts for everyone who needs access. Any Google-backed work account is fine; personal Gmail accounts are not.
4. An active Zus data agreement. Your access in BigQuery matches what you're already entitled to in Zus. If you use the Snowflake share today, you'll see the same data.
5. Somewhere to put query results in the same region. Your Zus dataset lives in the aws-us-east-1 region. If you save results to a table, that table needs to be in the same region, or Google charges you to move the data. More on this under cost.

What it costs and who pays

• You pay for your own queries. BigQuery bills query compute to the project that runs the query, not the one that owns the data. You run queries from your project, so they land on your invoice. You can't run jobs in the Zus project, so there's no way for your queries to end up on our bill or ours on yours.
• The rate is higher than standard BigQuery. Because the data is read in place in AWS, Google bills it at their cross-cloud rate: $9.125 per TiB scanned, against $6.25 per TiB in a normal US region - about 46% more. Check https://cloud.google.com/bigquery/pricing#bqomni for the current number.
• Watch out for one thing that can surprise you. If you join Zus data to a table of your own that lives in a normal Google region, Google quietly copies the data across clouds to do it and charges $0.09 per GiB for the move. To avoid it, keep the tables you join against in aws-us-east-1. On a small lookup table this is pennies; on a large fact table it isn't. Those cross-cloud joins can also run slower because of the transfer latency.

Limits and things to know

• Region. aws-us-east-1 is the only region available today. Despite the name, it's a BigQuery location like europe-west1, not something you set up on AWS. Google runs the BigQuery compute inside AWS data centers there, which is how they take on the AWS cost themselves. In practice: set the region picker in the BigQuery console to aws-us-east-1 or you won't see the dataset, and every query runs in that region.
• Deleted records are already filtered out. You won't see them, and you don't need to filter on any operation or status column yourself.
• Column types can differ from Snowflake. The same table can come back with slightly different types than the Snowflake share, because BigQuery reads the underlying file types directly. If you're moving a query across, check your casts.
• Adding or removing people. You manage in your Google Group.

Reference links

• BigQuery Omni pricing
• BigQuery pricing - on-demand compute
• BigQuery cross cloud transfer charges
• BigQuery custom query quotas
• BigQuery query permissions