Snowflake rolled out updates to its cloud-focused data platform that enhance what the vendor has been evolving over the past several years and also maintains its notion that a proprietary system is a better deal for customers.

Most of the updates build on the launch of past private previews. One of those updates is to its Snowpark service, which is inching closer to general availability. The platform allows developers to program data in their languages of choice and then execute extract, load, and transform (ELT) and extract, transform, and load (ETL) data modeling, data preparation, and analytics on Snowflake. This offers users a single platform for building data exchanges to easily share governed data and simplify an organization’s IT architecture.

“What is new now is that Snowpark is real,” Christian Kleinerman, SVP of product at Snowflake, explained during a press briefing, adding that the company has been adding customers to the platform since that initial unveiling last year. “It is real, not just because we say so but we have a fantastic lineup of partners delivering solutions for our mutual customers on Snowpark.”

Kleinerman added that the platform’s GA launch timing is “a function of customer feedback. We don't like to be in a period for too long, it could be three months could be six months, but we don't have a preset schedule.”

What did get a GA launch is Snowflake’s Data Marketplace. The platform allows users to create live, bi-directional access to data sets for a service provider to deliver endpoint business logics such as risk assessments or behavioral scoring.

Snowflake is also working on a “try before you buy” option for the Marketplace that will allow users to access sample data to see if it meets their needs before paying to access the service.

Kleinerman did note that customers will get the best marketplace experience if all parties involved are using the Snowflake platform, but he also said that it’s not required.

“The data sharing capabilities, which is the underlying technology of the data marketplace, does assume that there's a Snowflake endpoint on either side,” Kleinerman said, adding that they can also access the service if they use what he said are called “reader” accounts. These allow providers to establish a proxy endpoint on “behalf of non-Snowflake users.”

“We see that use on a regular basis as a stepping-stone and then you see customers and reader accounts going and graduating to their own small account and then actually it's a fairly successful motion for us,” he said.

Kleinerman added that this model is superior to what can be provided from open source platforms that continue to generate significant financial investment in the market.

“We believe that the experience that we can deliver by having snowflake on both sides is superior, dramatically superior to any alternative where an open protocol is the interface,” Kleinerman said. He cited comments from CEO Frank Slootman regarding specific use cases where users are also sharing “business logic through shared functions, and that is very difficult to do, especially with the privacy implications with an open protocol.”

“We look at what are the tools and approaches that we have at our service, and we optimize this always for best experience, best quality of experience, [and] minimal friction for our customers,” Kleinerman added. “And we believe that our current approach on data sharing and function sharing leads to the best experience for our customers. [It’s] not clear to us that an open protocol gets you there.”

Snowflake has found success with its model. A recent Forrester Wave ranking of cloud data warehouse providers had the vendor in the "leaders" category alongside rivals Amazon Web Services (AWS), Google, Teradata, Oracle, and Microsoft.

"Customers like Snowflake’s separation of storage from compute, as well as its scalability, time travel feature, ease of administration, data sharing across multiple lines of business, and consistent performance," the report noted.