Snowflake has announced significant advancements to its unified platform at the Snowflake Summit 2024, aimed at providing organisations with greater flexibility and interoperability in managing their enterprise data. The enhancements intend to simplify the discovery and collaboration on data, models, and applications, ultimately boosting performance and efficiency within the AI Data Cloud.
"Snowflake is making customers' data, models, and applications even more powerful by embracing open data and interoperability across the ecosystem," stated Prasanna Krishnan, Head of Collaboration and Snowflake Horizon at Snowflake. "We're providing customers with new ways to seamlessly access, understand, protect, and drive value with their data at the speed and scale they need to be successful."
The announcement highlights the introduction of Iceberg Tables, which now allow customers to leverage Apache Iceberg's open table format to enhance their data lakehouses, data lakes, and data meshes. This innovation is set to improve the flexibility and support for data stored externally, benefiting organisations such as Booking.com, Capital One, and Komodo Health.
Thomas Davey, Chief Data Officer at Booking.com, remarked, "Apache Iceberg’s large and diverse ecosystem of contributors and products made it a clear choice for us. With Iceberg, we can broaden our use cases for Snowflake as our open data lakehouse for machine learning, AI, business intelligence, and geospatial analysis, even for data stored externally."
The company has also rolled out the Polaris Catalog, an open catalog implementation for Apache Iceberg, enabling new levels of choice, flexibility, and control over data. Organisations can opt to run Polaris Catalog hosted in Snowflake's AI Data Cloud or host it independently using containers.
In addition, Snowflake Horizon, the company's built-in governance and discovery solution, has been enhanced to manage the complexities arising from the rise of AI. New features like the Internal Marketplace allow users to curate and publish data products exclusively for internal use while limiting unintended external access. Scott Richardson, CIO of Enterprise Data, Analytics & AI at Ally Financial, commented, "Snowflake Horizon's built-in governance, discovery, and protection capabilities ensure that we're operating with the highest degree of compliance, security, privacy, interoperability, and access."
Snowflake has also implemented Universal Search, enabling customers to perform natural language searches across the AI Data Cloud, Snowflake storage, external Iceberg storage, and third-party providers. This feature builds on technology from Neeva, which Snowflake acquired in May 2023. To further enhance utility, the platform will introduce AI-Powered Object Descriptions to generate relevant context and comments for tables and views.
The Snowflake Performance Index (SPI), which measures the platform's performance metrics, indicates a 27% reduction in query duration across stable customer workloads since its inception. Additionally, improvements have been made to data loading speeds, with up to 25% performance enhancement for JSON files and up to 50% for Parquet files.
Snowflake is expanding its AI Data Cloud to include highly regulated and sovereign markets globally. This expansion includes an EU-specific data boundary to meet European regulatory requirements and a dedicated environment for the Department of Defense (DoD) customers with networking integration meeting Impact Level 4 (IL4) security controls.
Other announcements at the summit included advancements to Snowflake Cortex AI and Snowflake ML, new tools to aid developers, and a collaboration with NVIDIA to build customised AI data applications within Snowflake.
