Snowpark Allows Organizations to Securely Run AI in Containers
According to a Snowflake press release, “containers have emerged as the modern approach to package code in any language to ensure portability and consistency across environments, especially for sophisticated AI/ML models and full-stack data-intensive apps.”
The company’s announced enhancements enable organizations to accelerate the development and creation of workloads that previously would have taken an enormous amount of time and compute power to manage. These workloads were further complicated because they couldn’t be run directly where data is governed. This “makes managing the security of the data being used in any of those scenarios very hard, and sometimes close to impossible,” Snowflake noted.
These limitations led the company to develop Snowpark Container Services, which will enable “developers to deploy, manage and scale containerized workloads (jobs, services, service functions) using secure Snowflake-managed infrastructure with configurable hardware options, such as GPUs.”
EXPLORE: Find out what your organization should know about containerization.
Torsten Grabs, senior director of product management at Snowflake, called the new container services the most exciting announcement of this year’s summit.
“It’s just such a fundamental capability that we are releasing with that, and it’s opening so many doors for us and our customers to create more value with their data.”
“We’ve been talking about bringing the work to your data for a very, very long time. But there were all of these little reasons that motivated customers to not do that and, rather, create data silos somewhere else. I think those days are over now,” Grabs said.
Snowflake’s Container Services Can Help Reduce Complexity
“Complexity is really harmful for organizations because it slows everyone down, it creates risk and creates additional cost. But our philosophy has been to provide one single integrated product experience across all of the various capabilities that we offer,” Grabs explained.
That philosophy extends to the way Snowflake thinks about containerized compute.
“We want to make this as simple and straightforward as possible, and then help our customers to have fewer and fewer silos in their data stack,” Grabs said. “And by virtue of having fewer silos, reduce complexity for their overall data architecture.”
“This single product philosophy actually carries forward in the way we’re thinking about security for our containerized, compute offerings,” he continued. “The same role-based access control that you have — that we literally apply to all of our concepts — that also applies to Snowpark Container Services.”
The offering is designed to help minimize complexity by eliminating the numerous security policies applicable in an environment employing multiple platforms. “This is exactly what we would like to avoid, as we want to have the same consistent security governance and privacy philosophy across the board and apply the same concepts everywhere,” Grabs said.
