Last week at the Semantic Technology and Business Conference in New York City, attendees got a chance to learn – or perhaps be reminded – of the business value of semantic technology at a host of keynotes and presentations. One of them specifically on that topic was hosted by Christopher A. Moran, CTO of consulting company IMSC, which provides information management services to clients in the DoD, civilian and intelligence agencies.
Some takeaways from the discussion follow:
- With semantic technology, extending what the organization knows about something – its customer accounts, for instance – doesn’t have to result in a big system overhaul. The added value the business gets comes as “simply an extension to the [semantic] graph,” as Moran termed it. Each system need only be designed to use the graph database.
- In the non-semantic world, data relationships are implicit in and data is fitted to a pre-engineered structure. In the semantic world, relationships are explicit and not tied to any one structure. As an example of why this matters, Moran posed the question of whether at the DoD, a group could know a year from now what its mission will be and what data it will need. The same, though, holds for any organization: “Does anyone know precisely what they will sell a year from now or what they are going to need?” he noted. “The only thing we can design a structure for is information we know we needed six months or a year ago. It’s like the saying about always fighting the last war. It’s the same in IT. We answer the questions we knew we had six months ago, not a year from now.”
- With semantic technology, the meaning of data is expressed by the data itself. Moran recalled a project he was working on some years ago, where the original database administrator who understood everything about all the databases from the company’s early days was not happy about being asked to add a new column to a system, and expressed his displeasure with some physical gestures. Throwing erasers, it seems, was involved.
Without semantic technology, “if the only people that understand the meaning of the data are people like [that], then we have a problem,” Moran said. With semantic technology, an organization doesn't have to rely on them to understand the meaning, because it is explicit in the data itself. Information, then, can be understood by the whole enterprise, he noted, not just by certain people who may not always be feeling cooperative.
- With semantic technology, no business system is directly affected by changes to another system, and integrating multiple systems is more seamless. As an example, Moran asked the audience to think of a bank that acquires another bank, and then to consider all the redundant IT systems that acquisition likely brings with it. “Integrating them with the graph is a lot easier than trying to write point-to-point integrations with all the databases you have,” he said. That said, if an organization has already done the work of building those point to point-to-point integrations, there’s no reason to change that. The graph can sit atop the infrastructure, and integrate any new system. Yes, IT will still have dependencies to address and maintenance to deal with until some future date at which those legacy systems will be retired, but it also means change can happen at a pace that the organization can accept.
“Getting to this [future semantic state] doesn’t have to be painful,” he said. “You can do it one system at a time.”