# Spanner Omni preview brings a downloadable Spanner deployment option

> Google previews Spanner Omni for infrastructure outside its managed service, with a developer edition explicitly limited to non-production use.

Canonical URL: https://www.devobs.io/news/news-google-cloud-spanner-omni-preview/
By: Elias Brooks
Published: 2026-09-06T11:58:54.636Z
Updated: 2026-09-06T11:58:54.636Z
Event date: 2026-04-22
Section: Data

Google announced Spanner Omni in preview on April 22, introducing a downloadable deployment option for Spanner outside the managed Google Cloud service. The [announcement](https://cloud.google.com/blog/products/databases/introducing-spanner-omni) describes configurations on virtual machines, containers, and Kubernetes, including private data centers and other clouds.

## Portability changes the operating model

The available developer edition is explicitly intended for non-commercial, non-production development and testing. Google says it excludes enterprise security features and directs commercial early-access inquiries through a separate path. That distinction matters more for adoption planning than a broad claim of database portability.

The technical announcement describes replacements for Google-specific storage and timekeeping dependencies. Those changes make the deployment option possible; they do not remove the need to assess the environment where a distributed database will run.

For an engineering organization, the potential benefit is a more consistent application database across locations. The corresponding question is who will own installation, upgrades, capacity, backups, and incident recovery in each location.

## Test the boundary you intend to cross

A useful evaluation should choose one actual portability requirement. For example, can the same schema and transaction workload run in a local development environment and the intended hosted environment with understood differences?

A disaster-recovery evaluation needs a different test: replication or transfer procedures, failover criteria, acceptable data loss, and a rehearsed return to the primary site. Merely starting a second database is not a recovery plan.

Teams should begin with the developer edition’s permitted scope and a written comparison of deployment responsibilities. Production decisions should wait on the commercial offering’s applicable support, security, and availability terms rather than transferring assumptions from managed Spanner.

## Source references

- <https://cloud.google.com/blog/products/databases/introducing-spanner-omni>