News

Ipedo Boosts Data Virtualization App

• By Kurt Mackie
• September 25, 2007

Ipedo XIP 4.3 now includes an incremental caching capability that improves query performance. The company also added support for parameterized views, allowing users to set up a query template and pass additional parameters to that view. Finally, Ipedo expanded its support for the open source database MySQL Enterprise.

Ipedo's product is an enterprise information integration solution or data virtualization solution. It's typically used by large organizations with lots of distributed data sources, such as banks, insurance companies and high-tech manufacturing industries. Ipedo's customers include Hewlett Packard, Sun and British Telecom, among others.

The problem of dealing with distributed data sources, where it's costly or time-consuming to move data around, isn't new. Companies typically have used the same database across different locations or focused on a data warehousing solution, or they've built home-grown solutions that have become difficult to maintain over time, according to Tim Matthews, Ipedo's cofounder and vice president of marketing.

The basic idea with data virtualization is you can take multiple databases and make it look like one virtual database, Matthews said. Ipedo, which has been in business for five years, has seen the need for data virtualization solutions grow.

"It's just a reality that large companies have lots of data spread in lots of different kinds of databases and they have ever-emerging needs for recording," he said. "And to make things more interesting, companies have Web services either as data source inputs, or maybe they want their databases to have outputs as Web services. So this is why the whole idea of virtualizing your databases or data sources is really picking up speed."

Ipedo's XIP solution sounds like service-oriented architecture (SOA), but it's something else. It uses a data services layer to create a "view" or data virtualization. The view is created using a drag and drop modeling environment or via command line. The solution helps get databases into SOA.

"A really inefficient way to try and get your database into SOA is to write wrappers for every database you've got," Matthews explained. "You're doing a number of bad things there. You're taking on a lot of unnecessary effort. You're probably writing suboptimal SQL. You may be violating some data governance or security policies. So what we do is we describe a mediation layer. We actually create views to individual or several databases. We create a model where we dynamically generate the SQL and then we publish the model as a Web service."

The XIP solution fills in some knowledge gaps in the IT personnel structure. In particular, database administrators (DBAs) can use the solution to write queries in the ever-familiar SQL query language and have the solution automatically generate queries in the XQuery language. It also helps in situations where DBAs don't understand SOAs.

"People who are running databases or business intelligence applications don't necessarily understand SOA, don't necessarily understand XML and don't necessarily understand Web services," Matthews said. "That's why working with something like our mediation layer is perfect because you can have your DBAs create these models and they understand SQL. We take care of publishing them as Web services and you can publish them out to a registry, so they are available to SOA architects."

The solution also helps developers by enabling them to write good SQL statements that perform well, Matthews added. Users of Ipedo's solution are typically developers or DBAs or, in many cases, BI analysts.

In many cases, it's when people are moving to SOA that they realize they need a data strategy, Matthews said. Ipedo's customers may have hundreds or thousands of data sources to contend with, so they are looking for a strategy to help maintain all they've built up over the years.

Another trend, as reflected in Ipedo's announcement, is support for the open source MySQL Enterprise database among companies. Matthews said that a lot of companies are using MySQL as a database for their new applications. Companies can buy several open source MySQL databases for the price of one proprietary system.

"It's clear that a large number of companies, especially banks, are using open source in a big way," he said. "A number of banks even have open source common centers internally. A lot of banks know the limits of open source databases vs. Oracle and DB2 and they're willing to use these databases in their applications within their limits. I think they are very keenly aware where MySQL, and, to a lesser extent, PostgreSQL is going and they are keeping an eye on that."