Among the blizzard of new language support for Snowflake’s data warehouse this week was a plan to beef up its data marketplace, which promises users somewhere to buy, sell, and otherwise share data on the platform. The problem is that it is far from the only game in town.
SAP, by sheer coincidence, also launched a data market plan this week, a few weeks after Databricks announced its open standard for data sharing, hoping to get a more vendor-neutral approach to the thorny problem off the ground. The application giant and the data lake poster child join other tools for cracking the nut, including those peddled by long-time data integration specialist Informatica.
The point of Snowflake’s Data Marketplace is to make it easier to ingest third-party data, into the analytics environment, as it is all in Snowflake’s architecture. Last November, it announced third-party service providers would have the option to enrich data by running risk assessments, augmenting a data set with behavioural scoring, or “simply outsourcing the more advanced analysis” without having to move the data, Snowflake said at the time.
Now the marketplace has hit general availability and the $33bn-IPO company has released new features including a try-before-you-buy option to allow users to access and evaluate sample data. It also talks up new usage-based purchase options to let companies buy and sell data entirely on the Snowflake Data Marketplace, which the vendor says would streamline the process of purchasing third-party data.
Snowflake CEO Frank Slootman last week told us the marketplace was “very important to our overall strategy.”
Meanwhile, senior veep for product Christian Kleinerman explained that data providers once had to transact outside the platform when selling their wares. “Now with monetization options in the platform, they will be able to implement one of a number of business models: usage-based, query-based, compute time, or time-based. Customers are going to be able to browse the marketplace and assess the different data and purchase through the platform.”
Though Snowflake has made an impact among investors, it is relatively new to the enterprise market. Here, SAP has more than a 40-year history, with considerable chunks of business data stuff inside its enterprise applications. The German vendor is now promising to make business data more sharable via the SAP Data Warehouse Cloud, the public cloud system.
Due to go live later this year, the plan is that customers will be able to access external data assets “in a matter of clicks that could take weeks or months today when integrating themselves,” SAP said.
Jan Gilg, president SAP S/4HANA, told us this week that the move would also allow partners to offer “pre-built content” for data-driven applications for their customer. “[T]he new data marketplace for data is a game-changer for partners because it’s so easy, it’s like the Spotify for data sharing.”
Coming at the data-sharing problem from a different angle again earlier this month, Databricks launched an open-source project called Delta Sharing, which will be donated to the Linux Foundation. Databricks said the open protocol would support secure sharing of data across organisations in real time, crucially independent of the platform on which the data resides. The initiative is supported by AWS, Google Cloud, and BI and visualisation firm Tableau.
At the time, Databricks CEO Ali Ghodsi told The Register that financial data providers such as FactSet, the New York Stock Exchange, and S&P Global had agreed to be part of the open sharing project, as had other data system providers like Starburst. “For the first time, you could produce a data set in Databricks, and you can share it with another company, which might not have Databricks, they might have Starburst. They can access the data securely, it gets tracked and audited.
“The problem with Snowflake is everybody has to buy Snowflake for [its marketplace] to work. The truth is a very small fraction of the data in the world is in Snowflake. Even in the cloud, they have a very small fraction of the data.”
But Doug Henschen, veep and principal analyst with Constellation Research, told The Reg data sharing was more than a simple matter of having an open data format. “What I like about the Snowflake Marketplace is that it’s clearly addressing the challenge of facilitating the exchange and monetisation of data. There’s real value in managing the marketplace, facilitating transactions and handling the invoicing, billing and payments for both parties.”
But another senior analyst, who asked not to be named, was much more sceptical about Snowflake’s promise. “The idea that everyone’s going to put everything into Snowflake is a nice aspiration for them to have. But it is one Amazon, Google, and Microsoft share with their various data platforms.”
The Databricks approach is more likely to appeal to data providers who are hoping to avoid lock-in to a specific vendor, he said. “The danger is you can make yourself dependent on a supply of data, and then you also have to use a particular vendor. People will want the choice; they don’t want to have to come through a particular product or channel.”
At the same time, Snowflake’s lack of maturity in the enterprise market might hold it back as the de facto data-sharing platform, he said.
“Snowflake is a hot product at the moment, and if it can get over the threshold where a network effect kicks in, then it might work, but it is not obvious. Snowflake is not that big a company; it is an order of magnitude smaller than AWS, for example, which has its own data sharing products. Meanwhile, SAP has the advantage that so many businesses are already using its applications.”
Snowflake might be hot on the stock market, but cool heads should take a look at its enterprise track record before committing to its data-sharing platform. ®