Jennifer Zaino
SemanticWeb.com Contributor

The semantic web has the potential of redefining the way enterprises manage IT. That's the perspective of Stephen Lahanas, principal consultant and co-founder of Semantic Inc., a technology consulting company that specializes in enterprise solutions integration.

Lahanas cut his teeth in the system-integration trenches. With a background over the past 15 years working on system integration and enterprise transformation efforts for large projects -- including a multi-billion dollar effort by the Air Force originally conceived to integrate 1,000 systems and multiple data centers -- he's seen what doesn't work, and the possibilities of what could work.

The history of dealing with the consolidation or replacement of hundreds, thousands, or even hundreds of thousands of systems, with thousands more data interfaces, and trying to blend them into some coherent architecture has always run up against the problem that the technology wasn't really geared to make that happen in an easy fashion -- if at all. There have been plenty of tools and specializations aimed at solving it, "but it's still incredibly expensive and difficult to manage and often doesn't work," Lahanas says. Problems are exacerbated by the fact that in mega-projects, individuals tend to mole down into their silos of specialties and lose sight of the bigger picture, so their efforts never are linked together effectively.

"So what mechanisms could you use to link the pieces of this puzzle together? That's where we start looking at things like ontologies, taxonomies, RDF data stores, and the whole concept of communities of interest as a business mechanism that lets people define the semantics of their business," he says. "Those semantics are the business processes, the definitions of their terminology which become the data sources."

In the last few years, tools have emerged that make possible building ontologies and taxonomies in a visual manner, and exchanging them on a data-driven basis. "You must be able to view all these semantic structures as more or less pattern templates that can be compared and mediated or reconciled with one another as necessary," he says.

IT has always operated in a dynamic context, and that only become more true as new computing models and platforms take root. That just increases the odds that by the time a large integration project actually completes using more traditional tools and methodologies, it's already obsolete.

"The only way around that is to dynamically change a solution without recoding all the pieces that are interrelated," he says, and the only way to do that is through abstracted semantic mediation layers that connect logic, leaving the more complex elements like the database management systems and so on relatively untouched. Think of it as an extension of the concept of building interfaces on top of a core set of data tables. You don't want to change the data table logic all the time when all you want to do is provide a different set of views to customers the way they need to see it, he explains.

The biggest barrier to moving in this direction isn't the tool as much as it is a philosophical barrier, Lahanas says.

"We still don't understand that every time we build something new and different we add to our complexity," he says. Data models won't be the fabric that connects everything to mediate against that, as they don't help with service issues, with application logic issues, with network operations infrastructure issues. He harkens back to a project he worked on a couple of years ago, where over the course of a year or two the data model changed so radically from previous versions that anyone building their data structures based on that model would have had to recode them all from scratch. As the model grew, so did its complexity and the difficulty of managing it. "Once you get into that situation you lose your flexibility," he says.

Applying semantics to enterprise architecture enables the connections and the flexibility. He sees immediate value to using semantic technologies for program lifecycle management, services-oriented integration, dynamic learning orchestration, and even cybersecurity.