New Snowflake service enables secure AI, ML deployment
Snowflake on Thursday made Snowpark Container Services generally available to enable customers to securely deploy and manage models and applications — including generative AI — within the vendor’s environment.
First unveiled in preview in June 2023, Snowpark Container Services is a fully managed service now available in all AWS commercial regions and in public preview in all Azure commercial regions.
Containers are a type of software that can be used to isolate applications for secure deployment. Snowflake’s new feature enables customers to use containers to manage and deploy any type of model, but they are optimal for generative AI applications because they enable customers to safely join large language models (LLMs) and other generative AI-powered tools with their data, according to Jeff Hollan, Snowflake’s head of applications and developer platform.
Given its role in helping user develop AI tools, Snowpark Container Services’ launch builds on Snowflake’s recent moves to customers with an environment for developing generative AI models and applications, according to Mike Leone, an analyst at TechTarget’s Enterprise Strategy Group.
Sridhar Ramaswamy took over as Snowflake’s CEO in February when Frank Slootman stepped down after five years of leadership that included guiding Snowflake through a record-setting initial public stock offering. Since then, the vendor has aggressively added generative AI-related capabilities such as launching its own LLM, integrating with Mistral AI and tools to enable users to quickly create AI chatbots.
“There has definitely been a concerted effort to enhance Snowflake’s capabilities and presence in the market when it comes to AI and more recently, GenAI,” Leone said. “Offerings like Snowpark are helping AI stakeholders like data scientists and developers use the languages they prefer.”
As a result of what it adds, Snowpark Container Services is a significant new feature for Snowflake customers, he continued.
“It’s a big deal for the Snowflake ecosystem,” Leone said. “By being able to easily deploy and manage containers right in the Snowflake platform, it helps customers more easily handle complex workloads and keeps things consistent across development and production stages.”
While Snowflake Container Services provides developers a secure environment, it was revealed in May that the vendor’s platform was nevertheless infiltrated by threat actors who used customer log-in credentials to gain access.
Among those affected by the security breach are AT&T — which filed a report with the U.S. Securities and Exchange Commission detailing the effects of the breach — and Ticketmaster. In all, over 160 Snowflake customers were reportedly compromised.
New capabilities
Generative AI has the potential to transform business by enabling almost any employee to easily work with data to inform decisions and by making trained experts more efficient in their roles.
Generative AI, when combined with an enterprise’s proprietary data, lets users work with that data using true natural language, greatly reducing the coding requirements and data literacy training previously needed to work with complex data management and analytics systems. As a result, non-technical workers can query and analyze data while trained users such as data engineers and data scientists are freed from mundane tasks that occupy much of their time.
With generative AI’s potential, many data management and analytics vendors have made developing generative AI-powered features a focus of their product development.
In addition, many enterprises have started to build models and applications trained on their proprietary data so the models and applications understand their business and can help inform decisions.
Among data platform vendors, AWS, Databricks, Google, IBM, Microsoft and Oracle have all focused on providing customers with environments for developing generative AI tools. Snowflake was perhaps not as aggressively developing its own environment for generative AI development as its rivals while Slootman was CEO.
Now, the vendor appears as committed as its competitors. But it hasn’t yet developed an ecosystem for generative AI development that quite matches those of its rivals, according to Leone.
“Snowflake has gone as far as creating their own LLM,” he said. But they still have a way to go to catch up to some of their top competitors.”
Matt Aslett, an analyst at ISG’s Ventana Research similarly said Snowflake is catching up to its rivals.
The vendor initially focused on traditional data warehouse capabilities, he noted. In late 2023, however, Snowflake unveiled Cortex, a platform for developing AI models and applications The platform includes access to a variety of LLMs and vector search capabilities and, even though many Cortex capabilities are still in preview, marked a significant step forward, according to Aslett.
“While rival providers were in some cases quicker to add support for GenAI models and development, Snowflake took a leap forward in late 2023 with the launch of Cortex AI,” he said.
The general availability of Snowpark Container Services furthers Snowflake’s attempt to foster generative AI development.
The feature provides users with on-demand GPUs and CPUs to run any code, written in any language, next to their data, according to Snowflake’s Hollan. That could enable them to deploy and manage any type of model or application, whether AI or not. But it does include AI.
And critically, due to the containerized environment, Snowflake customers never need to move their data out of the vendor’s platform when developing generative AI, machine learning and other applications.
“It’s optimized for next generation data and AI applications by pushing that logic to the data,” Hollan said. “This means customers can now easily and securely deploy everything from [source code] for their application to homegrown models in Snowflake.”
Beyond providing a secure environment, Snowpark Container Services aims to both simplify model management and deployment and reduce the cost typically associated with model oversight.
Deployment and management often requires piecing together various services from different vendors. Snowflake removes that need by providing a fully integrated managed service. In addition, by fully managing the containerized environment and adding a budget control feature, Snowflake says it is reducing operational costs while providing a measure of cost certainty.
Beyond its containerized environment and cost control features, Snowpark Container Services includes the following:
- Diverse storage options to support applications such as deploying LLMs including local volumes, memory, Snowflake stages and configurable block storage.
- Snowflake Trail to provide observability through integrations with observability specialists such as Datadog, Grafana and Monte Carlo.
- Streamlined DevOps capabilities including programmatic ingress to help automate software development and IT operation tasks.
Aslett noted that Snowflake’s launch of the Snowpark development environment in 2020 was an important means of enabling data engineers, data scientists and developers to use the coding language of their choice in concert with their data.
“Snowpark Container Services takes that a step further,” Aslett said.
The feature advances Snowpark by letting customers use third-party software such as generative AI models and data science libraries with their data, he continued. The managed service includes products and services from Snowflake partners including Amplitude, Dataiku, Nvidia, Pinecone and SAS — which enables users to leave their data in Snowflake for processing.
“This potentially reduces complexity and infrastructure resource requirements,” Aslett said.
While now generally available, it took Snowflake more than a year to move Snowpark Container Services from private preview to general availability.
Between June 2023 and now, the vendor focused on advancing the feature’s governance, networking, usability, storage, observability, and development operations capabilities, according to Hollan. In addition, Snowflake worked to make it more scalable and improve its performance.
One Snowflake customer that used Snowpark Container Services during its preview phases is Landing AI, a 2017 startup founded by Andrew Ng that specializes in computer vision. Using the feature, Landing AI developed LandingLens, an application that trains and deploys computer vision models.
“[With Snowflake], we are increasing access … to AI for more companies and use cases, especially given the rapid growth of unstructured data in our increasingly digital world,” Landing AI chief operating officer Dan Maloney said in a statement provided to TechTarget.
Plans
With Snowpark Container Services now generally available on AWS, Snowflake plans to make the feature available on all cloud platforms, according to Hollan.
In addition, part of the vendor’s roadmap is focused on continuing to improve Snowpark Container Services with more enterprise-grade tools just as it worked to improve the feature throughout the preview process.
“Our team is investing in making it easy for companies ranging from startups to enterprises to not just build but also deliver, distribute, and monetize next-generation AI products across their ecosystems,” Hollan said.
Aslett said making Snowpark Container services available on Azure and Google Cloud is the logical next step. He noted that the release of the managed service is “significant” but that it needs to be more widely available than in just AWS regions.
“The next step will be to bring Snowpark Container Services to general availability on other cloud platforms and demonstrate to potential customers that Snowpark Container Services is [equal] in terms of security and reliability — as well as integration — to other Snowflake applications and features.”
Eric Avidon is a senior news writer for TechTarget Editorial and a journalist with more than 25 years of experience. He covers analytics and data management.
#Snowflake #service #enables #secure #deployment