Every enterprise navigates two distinct realms. One realm is characterized by applications that process orders, track events, and engage customers in real-time. The other realm encompasses analytics platforms that extract insights, train models, and drive artificial intelligence. Between these two worlds lies a complex web of extract, transform, load (ETL) pipelines, batch jobs, and third-party tools, all of which can impose significant maintenance costs.
Snowflake is poised to bridge this divide with the introduction of Snowflake Postgres. During the recent Snowflake Summit, the company unveiled two pivotal features:
- Data mirroring: An always-on replication feature between Postgres and Snowflake, set to enter public preview soon.
- Postgres for your data lake: A more adaptable method for synchronizing Postgres with analytics using open formats like Iceberg, which will be generally available shortly.
These innovations facilitate a seamless connection between transactional and analytical data, eliminating the need for complex pipelines.
Tackling the No. 1 infrastructure problem
Feedback from customers consistently highlights that transferring data between online transaction processing (OLTP) and online analytical processing (OLAP) databases ranks as the most challenging infrastructure task within their data ecosystems. The visible costs associated with this task—such as ETL licensing, pipeline compute, and connector fees—represent merely the surface of a larger issue. Beneath these costs lie data inconsistencies, governance risks, and the engineering hours consumed by maintenance, all of which contribute to delayed decision-making due to outdated data.
In an age dominated by AI agents and real-time applications, this traditional approach often leaves organizations lagging behind. For instance, a fraud detection model may struggle to identify current threats if it relies on data processed in the previous night’s batch load, while a pricing engine may fail to optimize effectively with a six-hour data lag.
With Snowflake Postgres, a fundamentally different and radically simplified method for integrating Postgres and Snowflake has been developed.
Two new ways to connect your data: Always-on data mirroring and open-format data lake integration
Utilizing the open source pg_lake extension, users can now opt for either always-on data mirroring or flexible, open-format data lake integration to ensure a seamless connection between transactional and analytical data.
1. Data mirroring: ‘Set it and forget it’ data replication
Data mirroring offers low-latency replication between Postgres and Snowflake. Once a mirror is established, Snowflake automatically maintains target tables that reflect the current state of their source tables, accommodating schema changes and new tables created within mirrored schemas.
This setup can be completed in just a few clicks through Snowflake CoCo, the Snowsight UI, or a single SQL command. Once configured, your data flows effortlessly to where it needs to be. A demonstration of this functionality is available for viewing.
