Amazon began offering petabyte-scale data warehousing technology in February this year. Teradata has now responded with its own cloud-based data warehouse.
The move has not escaped the rest of the market. Teradata, which says it provides data warehousing technology to 90% of the world’s top banks, has announced its database, analytics platform and applications will also be available in the cloud, initially in the US.
While both firms could eliminate the hefty upfront capital investment needed to create a large-scale enterprise data warehouse, their approach to pricing differs. Amazon offers its data warehouse at $1,000 a terabyte per year, which it says is one tenth of the cost of most systems. However, the price of Teradata's cloud is based on a calculation of the total cost of ownership of equivalent in-house systems, or TCO neutral, as the company puts it.
Given Teradata now offers its Extreme Data Platform 1700 on premise for $2,000 per terabyte, Amazon's claim to be a tenth of the cost is less certain. Yet the perception among IT users and investors remains that cloud should be cheaper than in-house systems.
More on data warehouses
- Data warehouse concepts help build strong application architecture
- Microsoft SQL Server Parallel Data Warehouse (SQL Server PDW)
- Building data warehouses for mobile access is no small feat
- Integration between SAP Data Services, Business Warehouse on the rise
But those who build and run data warehouses will understand why this is not necessarily the case for their systems, says Tony Cosentino, vice president and research director at Ventana Research: “Generally, they are sophisticated people and they understand the drivers.
“TCO neutral is a reasonable position. Although cloud players market themselves as less expensive, there is no such thing as a free puppy. With cloud you still have a lot of expenses associated with it, the big difference is the way the cost is distributed.”
These hidden costs of cloud include the dedicated network infrastructure to connect to the data warehouse as well as data governance and compliance, which apply to both cloud and in-house systems, Cosentino says. “When people talk about Amazon, or other cloud services, they are not going to tell you what it’s going to cost for a T1 line to pin that up and get the data back and forth.”
He also says big suppliers can offer finance on large hardware investments that can make the amortised cost of an in-house data warehouse comparable to the running it in the cloud.
More on Teradata
Scott Gnau, president of Teradata Labs, says some customers may have the perception that cloud services should be cheaper than in-house systems. “There are a lot of companies that have a cloud mandate because it is a cool thing, or they think the cost is lower, and in general purpose computing that is largely the case. In high performance computing, where we fit, that is not the case.
“When you add in the cost of data transmission, data movement, dedicated communication links, all of the extended services and other things to make an offer meaningful, [cloud] can be a more expensive than doing it yourself.”
Meanwhile, if application data is physically close to the data warehouse it can be quicker to move data across, Gnau says.
Why look to the cloud for data warehousing and analytics?
Ventana's Cosentino says it will appeal to businesses that want to test or pilot data warehouse and analytics systems before jumping in with both feet. This flexibility will ultimately be the appeal of running these technologies in the cloud, he says.
Teradata also announced customers for its cloud offering. Online video distributor Netflix has selected Teradata Cloud for massively-parallel analytics.
However, do not expect Teradata stalwarts, such as Tesco and Sainsbury's, to move in-house systems to the cloud any time soon.
“Not a lot of people are wholesale moving the environments to the cloud. They are still keeping them on premise, especially large and mid-size companies that have legacy infrastructure," Ventana says.