The TL;DR
Tabular is a serverless, managed cloud storage platform built on Apache Iceberg. Iceberg is a popular open source table format for big data that provides data warehouse behavior on a data lake, and is in production at scale in data-rich companies like Netflix, Apple, Stripe and Pinterest. Tabular’s founders Ryan Blue and Daniel Weeks created Iceberg while at Netflix.
A key component of the Tabular platform is a catalog, which is fully compatible with the Iceberg REST protocol. Tabular provides the following services:
- Data loading – file ingestion, CDC for database mirroring and stream ingestion (coming soon)
- Intelligent storage optimization – regularly determining the best settings per table, rewrites data based on changes in the data profile (not a cron job).
- Centralized role-based access control – UI-based configuration, enforced per query.
- Concurrent access via multiple compute methods.
Tabular pricing is consumption-based and predictable. There is a free tier for smaller projects.
Tabular deploys as a SaaS offering, or a Tabular-managed service deployed into the customer’s cloud account.
Why do companies adopt Tabular?
- De-risks and speeds up bringing Iceberg to production
- Reduces storage and query costs
- Continuously ingests data into Iceberg
- Improves slow query performance
- Reduces data copying
- Unifies RBAC across compute methods
Where does Tabular fit in your data architecture?
- Iceberg table data and metadata resides in your account in the open Iceberg format.
- All data is secured by Tabular RBAC policies. Any engine can connect using the Iceberg REST catalog protocol.
- It runs background tasks to responsively optimize the file system as new data arrives and query patterns change.
- Tabular does not sit in the data path.
Notable Tabular users
Tabular customers span a number of industries, and include:
- A top 10 gaming company
- A top 20 financial technology company
- A top 10 real estate marketplace