Cohesity is using the Snowflake data warehouse’s external table feature to enable Snowflake analytics to access SmartFiles data.
Snowflake provides cloud data warehouse facilities running in AWS, Azure and GCP. Its external tables feature stores third-party file and object metadata and is a means of having Snowflake’s SQL analytic routines access data held outside Snowflake without having to migrate the data to Snowflake’s cloud stores.
Dell, MinIO and Pure Storage (FlashBlade) already have external table connector integration with Snowflake. A Snowflake analytic routine can, with external table support, access a MinIO object storage bucket as if it were a local Snowflake table store. Snowflake analysis is possible in instances when the data cannot be moved for compliance or other business reasons. Now Cohesity has joined in the external tables fun with its SmartFiles product.
Tarik Dwiek, Head of Technology Alliances at Snowflake, said in a statement: “We look forward to partnering with Cohesity to allow access to SmartFiles in the cloud or on-premises through Snowflake’s single, integrated platform.”
SmartFiles is a software product based on Cohesity’s scale-out Span File System which provides a way to store and access non-tier 1 file and object data in a Cohesity repository on-premises or in the AWS, Azure and Google clouds. It provides global deduplication, compression and small file optimization, claiming this lowers storage cost. SmartFiles can automatically tier little-accessed files and objects – cold data – to secondary storage on-premises or in the public cloud to save cost. This tiering of cold data follows analysis of data access trends on third-party tier 1 NAS storage with automated data movement to a cost-optimized tier.
In effect the SmartFiles store is a separate silo from Snowflake and the Cohesity-Snowflake external tables integration unifies the two silos. Cohesity says this gives you the freedom to manage and analyze your data wherever it resides while adhering to internal compliance and regulatory policies. It also means you can avoid the cost of migrating data into the Snowflake cloud.
Customers can accelerate query performance by creating materialized views on external tables.
Cohesity’s post says that, looking ahead, it has development efforts to generate AI-ready data that’s clean, accurate, and available wherever needed for AI routines to process it.
Snowflake competitor DataBricks also supports external tables. Dremio similarly allows external queries.