If you want to know where interest in Semantic Web technology stands as an adoptable technology, consider this: In March, the standards group OMG held a one-day information session in New York City called, “Demystifying Financial Services Semantics.” They expected about 50 to 60 people, but more than 200 showed up, according to Lee Feigenbaum, vice president of marketing and technology for Cambridge Semantics.
“These weren’t consultants or vendors,” Feigenbaum said. “These were people from the JP Morgans and Citi and Bank of America and Wells Fargos, et cetera, of the world that wanted to learn about semantics.”
It’s been about a year since I really talked with someone about the status of Semantic Web technology, so during a recent phone conversation, I asked Feigenbaum to provide a “state of Semantic Web technology.”
He explained it’s definitely moving out of the protocol stage and into full production, particularly in industries such as financial services and pharmaceuticals.
There’s often a lot of confusion about the term “semantic technology,” which can be used to mean everything from text analytics to search. Feigenbaum is very clear that’s not what he means.
He’s talking specifically about Semantic Web technologies, which refer to “a cohesive, coherent set of standards from the WCC.” These standards are being developed to create the next generation of the Web, but Semantic Web technologies actually are already benefiting enterprises.
In fact, for a long time, it’s been predicted that Semantic Web technology could change data integration.
It turns out Semantic Web technology absolutely is being used to make data integration more flexible, according to Feigenbaum.
First, it doesn’t require the rigid data models that you must have with relational databases. That makes it able for you to add queries for information in days or weeks, as opposed to months or years, he explains in this Q&A. Second, it uses a graph model that’s a much more flexible way to represent information than the traditional table-based model.
That also means you’re able to generate results from a project in a month or two, as opposed to the six- to 12-month timeframe you’d traditionally find with these types of data projects.
He shared an example of a pharmaceutical company that wanted to do a manufacturing lot genealogy — basically they wanted to know all the details on a product, from raw material through to the finished product and all the manufacturing and formulation variations in between.
“All this information was in different databases and different spreadsheets and random pieces of paper and all over the place,” he said.
The company obtained an estimate using traditional technology and was told it would take 9 to 12 months to do the project. That means the cost, just in terms of staffing, would be pretty high.
Using Semantic Web technology, Cambridge Semantics got it up and running within a week, he said.
“Not only could they start using it, but they could see whether it worked or not in a ridiculously short time scale compared to the 9 to 12 months that they had been quoted by IT,” he said. “So while the software itself is enterprise software and often has pricing to match, although probably not pricing at the level of an IBM or an Oracle for the most part, it can really minimize risk just by giving you results and letting you see whether something works or not really quickly.”
He added that Semantic Web technology could be added as an overlay, which means you can build from scratch or add Semantic Web functionality over top of your existing relational databases or other infrastructure. The solutions vary, from databases and individual tools to middleware platforms, which means the cost of entry varies widely as well, he added.