7 min Analytics

At Summit 2023, Snowflake expands the foundation for data apps

At Summit 2023, Snowflake expands the foundation for data apps

Snowflake is hosting the Summit 2023 conference this week, where it is unveiling innovations for developers and data experts. For these IT professionals, Snowflake aims to provide the foundation for data apps. SVP of Product Christian Kleinerman explains to us what features are coming to the Data Cloud in the near future.

Kleinerman points to the position of the Data Cloud within enterprise organizations today. “Very important for us is that Snowflake is a single product,” Kleinerman says. That’s something he says enterprises are looking for, if you compare it to how some tech vendors operate. They market separate products for a variety of use cases that might be subsumed into one product. That makes acquisition and use more difficult.

“A variety of workloads are possible because of the single integration, which reduces the overall cost of adopting Snowflake,” Kleinerman continued. He mentions things Snowflake is often associated with, such as data warehousing, data lakes and BI. Data engineering and cybersecurity, where there has been more focus recently, are also applications you can use Snowflake for.

The new features Snowflake is unveiling at Summit 2023 will need to support those applications a bit further. That’s why there are three innovation themes taking center stage at the conference. They are about Snowflake as a single platform, distributing applications and datasets and, finally, supporting programmability without losing security out of mind.

Building Apps with Native App Framework

What stands out as far as we are concerned when we look at how Snowflake fills it out is the public preview of the Native App Framework on AWS. This gives developers tools for building apps. “Every type of data application has historically required customers to move or copy their data and entrust it to third-party vendors, which is particularly problematic when customer data is highly sensitive. The Snowflake Native App Framework reimagines the status quo, enabling developers to bring their appsdirectly to their customer’s data, without that data ever leaving the customer’s environment,” Kleinerman explains. “We’re making it easier, faster, and more efficient for developers to build and ship leading apps at scale, so they can focus on earning revenue from some of the largest companies in the Data Cloud, with Snowflake seamlessly taking care of security, privacy, and governance concerns.”

Making the Native Application Framework available, due to its private preview since mid-last year, had been in the air for a while. The whole idea is that the Data Cloud makes it easy for enterprise organizations to centralize data. This creates a new model for application development and deployment, Snowflake claims, where developers can create multiple apps on a central copy of the data.

User-Defined Functions (UDFs) and stored procedures, for example, are available for building those apps. Companies can thus publish Native Apps in the Snowflake Marketplace. According to Snowflake, more than 25 apps are now available, while there are also more than 100 providers currently creating apps for different use cases. One example is Goldman Sachs, which built the Legend app using the Native App Framework. The app combines data platforms to create business insights for customers, business partners and engineers. To do this, it supports sharing datasets and connecting customers with their own data.

Text continues after the box below.

Experience of developers, data scientists and ML engineers

With the amount of announcements during the Snowflake Summit, the more focused functionalities are in danger of being underexposed. For example, for data scientists and ML engineers, the new Snowpark Model Registry is on the agenda. With this unified repository for models, Snowflake aims to further streamline MLOps. Snowpark Model Registry offers users a central place for model publishing and discovery, which should make it easier to bring models into production.

In addition, for data scientists and other Python developers, Snowflake is coming up with tighter integration with Streamlit. This open-source app framework is used to build and share data apps. Through the integration, the target audience should be able to “increase the impact of their work by building apps that bridge the gap between data and business action. With Streamlit in Snowflake, builders can use familiar Python code to develop their apps, transforming an idea into an enterprise-ready app with just a few lines of code, and then quickly deploy and share these apps securely in the Data Cloud.”

Furthermore, Snowflake will soon release a private preview of Iceberg Tables. With this, Snowflake is capitalizing on the popularity of Apache Iceberg as a standard for open table formats. Through Iceberg Tables, organizations can work with data in their own storage system in the Apache Iceberg format. This allows companies to deploy the Snowflake platform regardless of which IT system the data is managed in. “This simplifies data management by eliminating the need for organizations to move or copy data between systems,” Snowflake said.

More infrastructure and programming options

One announcement that ties in with Native App Framework is Snowpark Container Services. This feature, currently in private preview, is part of the Snowpark development framework. Data engineers use the framework to migrate pipelines, but it can also be useful, for example, for data scientists to build and train models. Snowpark Container Services should enable more workloads, by letting users use any programming language and support on more infratructured deployment options.

To build momentum for this, Snowflake announced (renewed) partnerships with various software and application vendors. These include the ability to “use popular AI platforms and ML features from Alteryx, Dataiku, and SAS to run more advanced AI and ML processing, and manage these data workflows with Astronomer’s platform powered by Apache Airflow — all entirely within Snowflake.” The image below provides an overview of the associated partnerships.

As you can see, Nvidia is one of the vendors Snowflake is renewing partnership with. It focuses on Nvidia accelerated computing and software integrations for Snowpark Container Services. This complements the two parties’ new collaboration to jointly bring generative AI capabilities to enterprises. In a seperate article, we discuss what this anouncement means.

What else is Snowflake doing on the LLM front?

With Nvidia news, Snowflake is paying attention to the large language model trend during Summit 2023. However, with Document AI (private preview), there is another update to bring the power of LLMs to the Data Cloud. This model supports extracting information from documents, such as invoice amounts and contractual terms, using natural language processing. Document AI is based on technology from Applica, a Polish company that Snowflake acquired late last year.

Snowflake’s main goal with Document AI is to capitalize on the expected explosion of unstructured data from documents, images, videos and audio. According to the company, extracting value from this data is still too complex. This is because it involves manual, error-prone tasks. Moreover, too few personnel with the right skills are available to actually extract value from this data, Snowflake states.

Foundation strengthened

Snowflake demonstrated its strong commitment to innovation with all the updates at Summit 2023. Developers, data scientists and data engineers can count on a host of new features that should ultimately benefit the entire business. In this way, Snowflake is becoming increasingly relevant at a time when data is taking on a more crucial role within companies.