Vertica Pushes Database into the Cloud

The database-as-a-service trend continues as Vertica Systems brings its columnar database to the cloud.

Vertica Systems is taking its columnar database into the cloud, joining a number of other vendors that have made similar moves. But the market for on-demand database services remains nascent, industry observers said.

Vertica Analytic Database Cloud Edition is on-demand version of the company's grid-enabled columnar database, hosted on's EC2 (Elastic Compute Cloud).

The idea is to create large, high-performance analytic data marts without upfront data center costs. The service will also help larger enterprises quickly set up data warehouses, Vertica Systems CEO Ralph Breslauer said.

"We believe there is a very large market for cloud-based data warehousing and analytics," the CEO said in an interview with eWEEK. "Many companies-especially younger ones in fields like biotech, hedge funds, and analytic software as a service-have terabytes of data they would like to analyze. And Vertica for the Cloud enables them do that by greatly minimizing the cost and expertise required to create a scalable, high-performance data warehouse database."

Almost all current cloud databases are row-stores designed to handle OLTP (online transaction processing) workloads, Breslauer explained. Vertica's cloud edition, however, is designed specifically for query-intensive analytic workloads. The combination of column-oriented storage and aggressive data compression dramatically reduces disk I/O, Breslauer said, and along with Vertica's grid architecture speeds up processing.

The company joins a number of other database vendors in the cloud, including EnterpriseDB, Microsoft, with its SimpleDB offering and Google with Bigtable.

But despite the growing number of vendors looking to take their databases into the cloud, concerns over issues such as security and cost mean that widespread enterprise adoption of such services is years away, said Gartner analyst Donald Feinberg.

Vertica's service, however, makes sense for third-party software developers who want to put a DBMS in place for either proof-of-concept projects or short-term projects without having to buy and manage hardware and software, Feinberg said.

Jasmine Noel, an analyst with Ptak, Noel & Associates, said only five years ago she would not have believed businesses would upload their corporate data to external service providers. But attitudes are changing, she said, as companies get comfortable with the idea of storing information in the cloud through the example of

"I can see some SMB [small and midsize business] and enterprise LOB managers using those types of services as [temporary] data marts for short-term analytical projects to get answers to a specific questions," Noel said. "How fast that service will pick up will depend on ease of use-the less data management and schema expertise needed to get value, the faster the market will grow."