4 years in the past, Databricks noticed great complexity within the information panorama: separate catalogs for every platform, siloed governance instruments throughout clouds, and no unified method to safe AI property. We pioneered Unified Governance by launching Unity Catalog, an open, versatile catalog layer to handle entry, lineage, auditing, and discovery throughout all information and AI property.
At this time, Unity Catalog has grow to be the muse of the Databricks Knowledge Intelligence Platform and the business’s solely unified governance answer for information and AI throughout codecs, clouds, and engines. From open information sharing to fine-grained safety and information governance, Unity Catalog helps organizations deliver context, management, and confidence to their information property.
At this yr’s Knowledge + AI Summit, we’re asserting main improvements throughout Unity Catalog, delivering one of the best catalog for Apache Iceberg™, new enterprise consumer experiences, and clever governance to guard delicate information and guarantee trusted information high quality at scale.
Right here’s what’s new.
The Greatest Catalog for Apache Iceberg™
Organizations adopting a lakehouse are sometimes pressured to decide on between Delta Lake and Apache Iceberg™. That selection creates synthetic silos: limiting entry to the information and AI instruments that groups can use, fragmenting governance, and locking metadata into format-specific catalogs.
Unity Catalog eliminates the necessity to decide on. Constructed on open requirements, Unity Catalog is the one unified catalog that works seamlessly throughout codecs, engines, and clouds, making it the muse of the open lakehouse. Over the previous yr, following the acquisition of Tabular, we’ve invested deeply in Apache Iceberg to increase this imaginative and prescient. We’re excited to announce:
- Full help for the Iceberg REST Catalog API, permitting exterior engines to learn (Usually Obtainable) and write (Public Preview) to Unity Catalog–managed Iceberg tables. This can be a main differentiator out there, eliminating format lock-in and enabling full interoperability unmatched by another answer.
- Iceberg managed tables are actually in Public Preview, delivering best-in-class worth and efficiency, liquid clustering, predictive optimization, and full integration with Databricks and throughout exterior engines, together with Trino, Snowflake, and Amazon EMR.
- Iceberg catalog federation is in Public Preview, enabling you to manipulate and question Iceberg tables managed in AWS Glue, Hive Metastore, and Snowflake Horizon with out copying information.
- Delta Sharing for Iceberg is now in Personal Preview, permitting you to share Unity Catalog tables and Delta tables with any recipient utilizing Delta Sharing and eat them in any consumer that helps the Iceberg REST Catalog API.
Collectively, these capabilities break down format silos and set Unity Catalog aside as the one catalog that delivers really open, unified governance and interoperability. Try our weblog on Iceberg help to be taught extra about these bulletins.
Increasing Unity Catalog to enterprise customers
Knowledge platforms shouldn’t cease on the technical consumer. Enterprise customers want a transparent, constant method to discover, belief, and work with information. Unity Catalog now provides a unified basis for enterprise context to bridge the hole between information and enterprise groups.
Unity Catalog Metrics: One semantic layer for all information and AI workloads
Inconsistent metric definitions throughout instruments and groups have lengthy triggered confusion, misalignment, and an absence of belief in information. Unity Catalog Metrics, now in Public Preview on AWS, Azure, and GCP and Usually Obtainable later this summer time, solves this by making enterprise metrics first-class property within the lakehouse. Not like metrics outlined solely within the BI layer, which restrict reuse and integration, defining metrics on the information layer makes enterprise semantics reusable throughout all workloads, like dashboards, AI fashions, and information engineering jobs. Unity Catalog Metrics are additionally totally addressable through SQL to make sure that everybody within the group can have the identical view of metrics, regardless of what instrument they select.
- Outline as soon as, use in all places: Create metrics as soon as in Unity Catalog and use them throughout AI/BI Dashboards, Genie, Notebooks, SQL, and Lakeflow jobs. Upcoming integrations will prolong help to BI instruments like Tableau, Hex, Sigma, ThoughtSpot, Omni and observability instruments like Anomalo and Monte Carlo.
- Ruled and auditable by default: Licensed metrics include auditing and lineage out of the field, enabling trusted, compliant insights throughout groups.
“Unity Catalog Metrics offers us a central place to outline enterprise KPIs and standardize semantics throughout groups, guaranteeing everybody works from the identical trusted definitions throughout dashboards, SQL, and AI purposes.”
— Richard Masters, Vice President, Knowledge & AI, Virgin Atlantic
“Unity Catalog Metrics represents an thrilling alternative for Tableau clients to leverage the worth of centralized governance with Databricks Unity Catalog. By way of our deep integration and increasing roadmap with Databricks, we’re thrilled to assist take away the friction for our clients in leveraging Databricks to outline their core enterprise metrics.”
— Nicolas Brisoux, Sr. Director Product Administration, Tableau
New curated discovery experiences with clever insights
To totally empower enterprise customers, you should make trusted information simple to search out, perceive, and use. Unity Catalog is extending its business-aware governance with a brand new Uncover expertise, now in Personal Preview, a curated inner market of licensed information merchandise organized by enterprise domains like Gross sales, Advertising and marketing, or Finance.
AI-powered suggestions and information steward curation assist floor the highest-value property, resembling metrics, dashboards, tables, AI brokers, and Genie areas which might be enriched with documentation, possession, and utilization insights. New clever indicators spotlight information high quality, utilization patterns, relationships, and certification standing, serving to customers rapidly assess belief and relevance. Plus, with Databricks Assistant inbuilt, customers can ask pure language questions and get clear, context-aware solutions based mostly on ruled metrics.
We’re additionally introducing new clever capabilities throughout Databricks to make information discovery simpler and extra intuitive, wherever customers work within the platform. Powered by Unity Catalog, these options assist groups discover trusted information sooner and perceive its context at a look.
- Domains (Coming quickly): Manage information by enterprise space to align discovery with the group’s operations.
- Certifications and Deprecation Tags (Beta): Sign information belief and enterprise relevance throughout datasets, metrics, and dashboards. Tagged property prominently show their standing in authoring surfaces just like the SQL editor, conserving information high quality indicators seen all through the consumer workflow. Certifications and deprecation tags can be found as part of Tag Insurance policies Beta.
- Request for Entry (Public Preview): To streamline supply, customers can immediately request information entry on to the asset.
Further superior governance capabilities now out there
Excessive-leverage governance with scalable, attribute-driven controls
Scaling information governance turns into more and more difficult as organizations develop, with extra customers, groups, and information property to handle. Static insurance policies and guide controls can’t sustain, resulting in governance gaps, safety dangers, and operational bottlenecks.
To handle these challenges, Unity Catalog now gives clever automation and versatile, scalable controls to categorise delicate information, implement coverage persistently, and speed up safe information entry throughout the lakehouse.
-
Attribute-based entry management (ABAC): Outline versatile entry insurance policies utilizing tags that may be utilized on the catalog, schema, or desk degree. ABAC is on the market in Beta for row and column-level safety on AWS, Azure, and GCP.
-
Tag insurance policies: Tag insurance policies implement a governance layer for the way tags are created, assigned, and used throughout Databricks. These account-level insurance policies guarantee tags stay constant and trusted, supporting the whole lot from information classification to value attribution. Tag insurance policies can be found in Beta on AWS, Azure, and GCP.
-
Knowledge classification: Intelligently detect and tag delicate information throughout Unity Catalog. New information is scanned inside 24 hours to routinely detect new PII, minimizing guide effort and permitting groups to remain on high of information entry. When used with ABAC, Knowledge classification routinely protects delicate information based mostly in your entry management insurance policies. Knowledge classification is on the market in Beta on AWS, Azure, and GCP.
“Implementing column masking throughout greater than 5,000 tables was an infinite guide effort. With ABAC, we’re in a position to apply constant insurance policies dynamically, drastically enhancing each pace and governance.”
— Ramesh Balasubramanyan, Databricks Admin, SAIF
“Databricks Knowledge Classification has been a game-changer in our information privateness and safety technique. Paired with ABAC, it allows us to routinely safe delicate information with out limiting the information that our analysts want. The most important profit has been pace, with automated classification and masking considerably lowering guide overhead, liberating up our resourcing and saving our staff numerous hours every week.”
— Mary Tesfay, Knowledge & Analytics Lead, Corp IT, Navitas
Automated information high quality monitoring at scale
Unity Catalog now intelligently detects and helps resolve information high quality points throughout all of your tables with information high quality monitoring, out there in beta on AWS, Azure, and GCP. Knowledge high quality monitoring checks freshness—how not too long ago information has been up to date—and completeness—whether or not information volumes are as anticipated—utilizing information intelligence throughout total schemas. Customers are in a position to perceive the well being of information at a look with well being indicators, whereas information homeowners can perceive the precedence of points based mostly on downstream lineage, uncover the foundation trigger, and set alerts utilizing built-in logging and dashboards.
Get began with Unity Catalog, the muse of Knowledge Intelligence
Unity Catalog continues evolving because the business’s solely unified governance layer, the muse for safe, clever, and business-aware information platforms. Whether or not you’re constructing AI brokers, delivering BI dashboards, or sharing information throughout organizations, Unity Catalog connects all of it by means of a single, open catalog.
To get began, observe the Unity Catalog guides for AWS, Azure, and GCP.
Watch the Knowledge + AI Summit 2025 keynote from Matei Zaharia, Co-founder and Chief Expertise Officer at Databricks, to be taught extra about these latest bulletins.
Register for Knowledge + AI Summit and discover the information and AI governance observe
