CopperEye aims to help customers downsize databases

A company led by former database-software vendor executives is touting enterprise search technology that it says can help companies reduce their reliance on expensive relational database systems.

CopperEye's twist? Its software enables customers to put static transaction data in an easily searchable flat-file format on networked-attached storage (NAS) servers rather than on relational database servers residing on pricier storage-area networks.

The British company's Greenwich software runs on a server, continually discovering structured data files that include business transactions and indexing them. Greenwich exposes the data in those files through Open Database Connectivity, and customers use CopperEye's Search software to make SQL queries.

"The founders started out developing what they thought would be the next great extract, transform and load [ETL] tool, and on the way to building that they actually uncovered a new way of indexing transactions," says CEO Kate Mitchell, who previously worked at IBM and Oracle. ETL tools let companies move data, reformat it and load it into a relational database for analysis.

Rather than using the B-tree indexing method common to relational databases, Mitchell says, CopperEye uses a proprietary indexing method that stores transaction data - stock trades, phone calls or Web purchases - in a flat-file system.

MessageLabs, an English provider of hosted e-mail scanning services, uses CopperEye's software for finding the e-mail transactions of 13,000 customers. Each day MessageLabs processes 6 billion rows of log file data and 200 million transactions using CopperEye's software.

"The problem we were trying to solve is to access billions of messages that are flowing through our infrastructure to help a single client locate a single message so they can understand where it is in the mail-message workflow," says Carmen Carey, COO for MessageLabs. "These transactions are static, so keeping them in a relational database was not the way to go."

Steve Duplessie, founder and senior analyst for the Enterprise Strategy Group, says: "The key is that CopperEye can take data out of the database and make the database faster and easier to manage. With CopperEye, you don't have to buy the most expensive storage gear and manage it with the most expensive administrators."

Duplessie says the cost savings on storage hardware alone by implementing CopperEye could be enormous. "No more big licenses for Oracle, no more giant server upgrades and no more expensive Tier 1 storage upgrades. Put 80% of the database on an inexpensive NAS box," he says.

CopperEye competes with data-warehousing and relational-database vendors IBM, Oracle and Netezza. Unlike Netezza's Performance Server, which combines a server, storage and relational database in one device, CopperEye's software is storage-system-agnostic.

CopperEye Greenwich and Search run on a Linux or Unix server that attaches to a Gigabit Ethernet network. The products are priced based on the number of rows of data under management. Pricing starts at $50,000. A typical installation could cost as much as $250,000.