ReleaseBytes
gcp Google Cloud Blog ·

Google Cloud Announces Borderless Lakehouse for Cross-Cloud Data Access

blogaiawsazuregcpdatabrickssnowflakepreviewengineergcp-bigquerygcp-cloud-storagegcp-spanner
feature

Google Cloud is enhancing its "borderless Lakehouse" architecture, built on Apache Iceberg, to enable AI agents and Gemini Enterprise to query data across AWS, Databricks, and Snowflake without moving it. This initiative introduces catalog federation and zero-copy integrations, offering zero-copy cross-cloud analytics, bidirectional interoperability, and unified governance to developers and architects. The new capabilities, including Cross-Cloud Interconnects and intelligent caching, aim to reduce data transfer costs and latency for in-place AI and ML workloads, with preview availability for key federation features.

  • →Borderless Lakehouse with Iceberg REST Catalog Federation
  • →Zero-Copy Data Integrations for SaaS Applications
  • →Intelligent Cross-Cloud Caching
  • →Google Cloud Data Agent Kit for Gemini Enterprise
  • →Cross-Cloud Interconnects for Reduced Egress Costs
Features (4) ›
  • Borderless Lakehouse with Iceberg REST Catalog Federation

    The borderless Lakehouse, built on Apache Iceberg, now supports catalog federation for AWS Glue, Databricks Unity, and Snowflake Horizon via the Iceberg REST catalog. This enables Gemini Enterprise and conversational agents to analyze data across multiple cloud platforms and on-premises systems without requiring data movement, with federation capabilities now in preview.

  • Zero-Copy Data Integrations for SaaS Applications

    Google Cloud introduces zero-copy data integrations with major SaaS applications like SAP, Salesforce, and Workday. This allows BigQuery to query live application data directly and enables these applications to run BigQuery's AI engines on their data in-place, unifying finance, HR, and customer data without complex ETL pipelines.

  • Intelligent Cross-Cloud Caching

    Intelligent cross-cloud caching is introduced to the borderless Lakehouse, securely storing remote data fragments within Google Cloud. This feature aims to eliminate repeated, costly data transfers for subsequent ad-hoc or BI queries, improving performance and reducing costs.

  • Google Cloud Data Agent Kit for Gemini Enterprise

    The Google Cloud Data Agent Kit and Conversational Analytics API allow users to build and publish custom data agents directly into Gemini Enterprise. These agents operate across the borderless Lakehouse, utilizing Model Context Protocol (MCP) tools for secure connections to BigQuery, Managed Spark, and Cloud Storage, enabling self-service analytics and grounded AI results.

Enhancements (2) ›
  • Cross-Cloud Interconnects for Reduced Egress Costs

    To address cross-cloud analytics challenges, Google Cloud is introducing Cross-Cloud Interconnects, offering private, dedicated links for consistent bandwidth and lower latency between clouds. These connections aim to reduce egress costs, particularly for accessing AWS data with zero variable egress costs, and support predictable monthly pricing with an SLA.

  • Knowledge Catalog for Universal AI Agent Context

    The Knowledge Catalog, an agentic context engine, synchronizes with AWS Glue, Databricks Unity Catalog, and Snowflake Horizon (all in preview) to provide a unified view of enterprise context across clouds. It translates raw schemas into business terminology and column-level lineage, enabling trustworthy AI decisions and automated governance for AI agents.

Read the original announcement →

https://cloud.google.com/blog/products/data-analytics/introducing-the-borderless-lakehouse/

Related releases