Tabular is a managed service that has a SaaS option and an AWS private cloud deployment option.

SaaS

In Tabular’s SaaS deployment, the catalog, table optimization, and data pipeline services all run in Tabular’s cloud accounts. Iceberg table data and metadata are read and processed in Tabular’s cloud account, but not stored there; output data is always written back to storage in the customer’s cloud account. Access requests to the data are controlled by Tabular’s role-based access control implementation and Tabular maintains an audit log of all authorization decisions.

Tabular-managed private cloud deployment on AWS

For a Tabular-managed cloud deployment, you provision a dedicated cloud account (that you own and control) into which Tabular is deployed. Data is read and processed in this account but not stored. Data is stored in your primary data account and only transits to the account provisioned for Tabular for optimization, after which the resulting data is written back to your primary data account.

Data and metadata never leave your accounts. Tabular delegates access to data files according to permissions set by RBAC policies in the Tabular catalog. This allows direct access to the data within your primary data account.

To manage your Tabular software, Tabular only needs access to the account assigned to Tabular. We require access to monitor, update, and remediate issues for our services, but do not require access to infrastructure to your other accounts. Nor do we require any connectivity between your accounts and Tabular-owned cloud infrastructure.

Regions

Tabular currently runs in AWS regions in North America, Europe, Asia and Africa. We can quickly add other regions based on demand.