Cloud-based data warehouse darling Snowflake has started its newest company in financial services, while Teradata, a kind of fixture in data warehousing for banks and insurance companies, is trying to increase its appeal with implementations of machine learning.
Although the world has focused on the stratospheric rise of Snowflake – the value rose from $ 1.5 billion.
The so-called cloud-native data warehouse business launched what is known as the Financial Services Data Cloud this week, accompanied by the claim that 57 percent of Fortune 500 companies in the sector are on its platform. It is described as an industry-specific platform that merges Snowflake technology with “partner-supplied solutions” and “industry-critical data sets”.
The idea is for financial services companies to use it to bring new customer-centric products to market, making the process easier by having all of the data already on the platform, the company said.
According to Snowflake, companies can leverage built-in security and governance to collaborate on data projects. Features include private connectivity for multiple public clouds, improved encryption with Bring Your Own Key (BYOK), integrated classification and anonymization of sensitive data, and integration with third-party tokens in accordance with SOX standards.
The platform offers the ability to securely share data across multiple public clouds, with support for sharing multi-tenant environments. A data catalog is to come from Alation. There are now a number of partner functions, including the Aladdin Data Cloud from investment management company BlackRock – a Snowflake-based system designed to help investment managers make better use of data – SI Cognizant, the data integration platform Dataiku, and Deloitte.
Matt Glickman, Snowflake-Veep, said the company’s platform would become the “go-to destination for traditional and alternative records” in financial services, while third-party technology companies were also helping customers build new services with the system.
He also claimed that customers in the financial services sector, working with service companies, are “putting more and more of their most important manufacturing tasks into a snowflake”.
This may come as a surprise to some cloud-wary observers. An insider with years of experience working with data warehousing systems in financial services said banks and insurers are struggling to find reasons to move their trusted on-prem systems to the cloud because security, performance, and costs are still issues be.
While Teradata has made a name for itself in the financial services space, where HSBC and Lloyds are its customers, Snowflake is new to supporting core workloads in this market.
The insider said that Snowflake can handle eight concurrent users per cluster, which means the cost increases as the system adds more users.
According to Teradata, “while the current system is known to have limited capacity in terms of throughput, compute power, and storage, it is guaranteed to be a known cost. It is the unpredictability of costs that is a huge shock to people.” [using Snowflake] because not only can they not explain the bill for this month, but they also have no idea what the bill will be for the next month, “he said.
In addition, technical teams in the financial services industry were skeptical that Snowflake had anything new to offer because it was based on a relational database. “There is no longer any leeway to advance the state of the art,” said the insider.
Although Snowflake had helped with onboarding and probably also with data exchange, it didn’t really solve a problem. “What is the technology enabler in this announcement, what is the new IP that is being brought into play?” he asked.
An analyst for a global technology research firm, who refused to be named, agreed that Snowflake was struggling to break into key data warehouse systems used by financial services companies.
Although they may perform tactical analysis and computing in Snowflake, banks and insurers rely on features Teradata has developed around its on-prem data warehouse appliances for 40 years.
“Teradata has built in machine learning [for performance] and integration of multiple data types: things Snowflake has not yet developed. If you have a large bank that relies on these features, migrating to Snowflake in the cloud is a bit tricky, and from a cost-benefit perspective, it’s hard to argue that Teradata now has a cloud system too. I don’t think they’re going to hit Teradata anytime soon, “he said.
Meanwhile, Teradata has announced new capabilities to bring machine learning models for platforms like the data lake spinner Databricks right into its own.
Scott Toborg, Teradata Director of Product Management, said The registry For example, a logistic regression model that a data scientist created in Apache Spark can be exported to the Predictive Model Mark-Up Language (PMML) interchange format, with coefficients, distortions, and other parameters used to describe that model.
“Then we import that file into Teradata, extract the data, and then build the necessary Java code to run this model again,” he said. The industry analyst said combining Databricks (Spark) to build machine learning models and Teradata to deliver them to business data could be challenging.
Although Snowflake has been in the spotlight of investors in recent years, insiders say it still has a long way to go in convincing core data warehouse users to migrate their main systems as it takes in peripheral and tactical workloads.
Established providers like Teradata have their own cloud history. For example, according to a customer report, Unilever has migrated its on-prem Teradata system to Azure in the past 18 months.
Meanwhile, Teradata is adding more features, meaning the battle in enterprise data warehousing is far from over. ®