News

Can XML solve data federation conundrum?

• By Rich Seeley
• July 30, 2003

Snapbridge Software, a start-up company based in San Diego, claims to have solved the problem of enterprise data federation with a combination of XML standardization and algorithms developed by its engineering team.

The emergence of XML makes it possible to provide the easy-to-implement enterprise data federation software that eluded data warehousing and data mart vendors in the 1990s, said Benjamin Chen, Snapbridge chairman and CTO.

The product he helped develop, Snapbridge FDX, is in limited release at “Fortune-class” companies in the financial, publishing, supply chain management and telecommunications industries, with an official launch set for this fall, he said.

While based on XML, Snapbridge FDX uses a “schema-free” architecture to eliminate the need for data modeling and speed implementation, Chen said.

“I think there are appropriate use cases for doing a master data model approach,” he said. “But the way we’ve attacked the problem allows us to go in immediately and solve the current problem. And because we’re schema-less you can add new data sources in real-time; you can make changes to existing systems. Our technology is very incremental.”

A development team that included pioneers in XML and Java made it possible to create a new approach to gathering information from disparate data sources, Chen said.

“We have a set of mathematical algorithms that understand how to inspect and interrogate each of the live instances of data themselves, be it a relational database, be it a structured document, we don’t care,” the CTO told XML Report. “What we do is we basically produce a master virtual index. What’s unique about this is this index is not a schema. It’s a GPS[-type] location of where the actual data itself lives.”

The Snapbridge virtual index technology focuses on simple meta data and a straightforward approach to gathering information, which Chen calls a “Straight Through Processing model for data federation.”

“We understand where the data is,” he said. “We understand some traditional easy meta stuff about that data. But essentially we have the facility for saying how to get to the data.”

Within the virtual index is a representation of the XML info set, Chen explained.

“We’ve got a true hierarchical view of these sets of data in one index, coupled with the fact that we have access to this data as XML,” he said.

Snapbridge FDX avoids the traditional XSLT process, which can result in performance problems in enterprise applications, Chen said. His development team has built a high-speed XML federation engine, which he calls “a huge piece of technology.”

“What this federation engine knows how to do is fire off multiple executions of style sheet transformations in real time,” he explained. “We’re completely standards-based in this regard, so not only do [we] use traditional software-based transform engines, we can use hardware-based accelerators. We can line up multiple transform engines side-by-side, fire them off like a microprocessor that’s doing advanced pipelining and then reassemble the results for output.”

Chen claims this technology has potential customers and partners knocking at his company’s door since the product went into limited release.

“In the past month and a half it’s been pretty overwhelming,” he said.

More information on the Snapbridge approach to data federation is available at www.snapbridge.com.