EII: A database of databases

The lifeblood of Merrill Lynch is the information that the investment bank has collected about its clients, trades, investment products and stock prices. Trouble is, that information is scattered across four separate databases, and brokers can't get a clear view of all the information they need for vital activities such as trade execution and risk management.

So the company uses enterprise information integration software to create a single view of the information located in all four of its databases. The software, from vendor MetaMatrix, creates a "virtual" data warehouse that apparently gives users the illusion that all the data is located in one place.

Enterprise Information Integration (EII) is designed to give companies real-time access to information that's stored across multiple applications, databases and data warehouses. It's an emerging technology -- Aberdeen Group puts current spending on EAI at less than $200m (£107m) a year -- but one that's growing fast.

What is Enterprise Information Integration? 1. A directory. This 'database about databases' contains information about data stored in the databases, and the information stored in each database's data dictionary and information about data shared across databases. 2. Conversion capability. An EII product must often convert incoming data from non-relational to relational format or from relational to object format. When updates to back-end databases are involved, the EII must also convert the common format back into the target database's format. 3. A database veneer. To the user, and if possible to the developer and administrator, the EII solution must look like a single database. Therefore, EII products must support standard SQL and querying transactions. 4. Front and back-end communications. The EII will have a method of communicating with front-end programs and with users. It should also be able to send transactions to, and receive results from, back-end databases.

Driving that growth is the entry of IBM into the information integration market. IBM's DB2 Information Integrator is the third generation of data integration products, which provide access to structured and unstructured content across multiple platforms. IBM's presence is seen as a positive factor by smaller, startup EII suppliers. "We're happy for IBM to spend the marketing dollars and for us to get the benefit," says David Penney, European managing director of EII startup MetaMatrix. "IBM is broadening the market for everyone."

Other new entrants to the market include SAP, which has licensed MetaMatrix' technology for its NetWeaver integration platform. In addition, Group 1 Software purchased Enosys in 2003 and continues to market its EII technology.

How does it work?
EII uses metadata to create a virtual database that provides a single view of information residing in multiple sources, explains Phillip Russom, an analyst with Giga Information Group. In a business environment, this could mean your call centre agents could submit a single query to the EII platform and get a result that incorporates relational data from a DB2 database, email from Lotus Notes, images from Documentum and spreadsheet data from Microsoft Excel.

"The key characteristic of EII is that it lets you access a view of data as if it were physically located in a single database, whereas it might be from a number of sources," says Russom. "The user only sees a single set of data because the integration is done transparently by the EII technology, along with all the security, data integrity and query optimisation."