carloscastilla - Fotolia

Google looks to poach workloads for its cloud data warehouse

Google Cloud wants to tap Teradata and Netezza's installed base with a new migration service for BigQuery. It may appeal most to new or smaller deployments.

Google hopes its latest cloud migration offer will entice Teradata and Netezza customers to switch to its BigQuery cloud data warehouse service, but such a move likely doesn't make sense for customers with the most massive and complex implementations.

The offer consists of four components. Customers will receive expert advice from Google Cloud staff or partner-led professional services, as well as funding for proof-of-concept projects. Google Cloud will also provide free training on BigQuery and related technologies via Qwiklabs, Coursera or its own online classes.

Third, customers can receive help from Google Cloud engineers on architectural design, in the form of "personalized deep-dive workshops at no additional cost," Google said in a blog post. Finally, while Google Cloud services staff are available to help with the cloud data warehouse migration, some "qualified" customers may receive financial support from Google Partners to offset the migration and BigQuery implementation costs.

Beyond human help, Google has also created a migration utility called BigQuery Data Transfer Service, which is now in beta.

Google's rationale for the migration offer is no surprise. The company claims on-premises data warehouses are too clunky, expensive, inflexible and hard to maintain, whereas BigQuery provides a highly scalable managed service that can save customers big money.

The BigQuery migration offer comes with some conditions. It's available only to existing Teradata and Netezza customers who agree to sign a contract for a 2,000-slot BigQuery allocation, which costs $40,000 per month. Storage costs are extra, at $0.20 per gigabyte per month, with the first 10 GB free.

Google Cloud aims at target-rich environment with BigQuery migration offer

Data warehousing workloads are a priority for cloud providers because they are strategically important to help modern companies optimize their operations, gain insights into customers and beat back competitors through large-scale data analysis.

Doug Henschen, analyst with Constellation ResearchDoug Henschen

Google is hardly the only company after Teradata and Netezza's installed bases. "AWS Redshift and Snowflake have been very aggressive," said Doug Henschen, an analyst at Constellation Research, which is based in Cupertino, Calif.

Despite how attractive Google's migration offer may look on paper, real-world issues loom, Henschen said.

"The real question for existing [data warehouse] customers is how much do they have invested in their legacy platform and what alternatives do they have?" he said.

It is far from trivial to move a major Teradata installation to a new platform, Henschen added. He recently spoke with a Teradata customer, a telco that manages 2 PB of data spread across more than 100,000 tables. The system gets queried 60 million times per day and is associated with 50 applications. "That is not going to be rebuilt," Henschen said.

For such large organizations, which represent a big chunk of Teradata's customer base, a cloud data warehouse migration is not economically feasible. Instead, Google and AWS are casting a wider net in shallower waters.

The sweet spot that the Redshifts and the Snowflakes and Google are going after are smaller deployments.
Doug HenschenAnalyst, Constellation Research

"The sweet spot that the Redshifts and the Snowflakes and Google are going after are smaller deployments. That's the big part of the market," Henschen said. "There are very few companies that are petabyte-class."

For customers that want to start from scratch, cloud data warehouses have a lot of appeal, Henschen added. Many companies are drawn to the platform due to Google's related analytics services, such as BigQuery ML, he said.

Teradata hasn't rested on its on-premises laurels, either. It offers IntelliCloud, a managed cloud data warehousing service, on its own infrastructure, as well as AWS and Microsoft Azure. IntelliCloud answers Google's knocks about the cost and operational headaches that are supposedly inherent to on-premises data warehouse deployments, Henschen said.

As for IBM, which is noticeably absent in Google's migration offer, it has Db2 Warehouse on Cloud, which is compatible with both Netezza and Oracle workloads. IBM also offers related migration services.

Overall, there is certainly a market for data warehousing in the cloud, evidenced by the success of Redshift, said Curt Monash, founder of Monash Research in Acton, Mass. Sometimes, customers' data originates in the cloud. In other cases, investigative and exploratory data warehouse workloads are spiky and benefit from elasticity, he said. Lastly, it's easier to put new workloads on new platforms than to migrate old workloads.

"The twist on this in the data warehousing case is that instead of a workload being strictly new, it can be greatly expanded," Monash said. "Typically, that's because you're doing certain kinds of analytics on certain data, and you increase the kinds of analytics, kinds of data, or both: e.g., ad hoc query to statistical analysis to neural networks."

Dig Deeper on Cloud app development and management

Data Center