New Architectures Help Make Databases Smarter
Breaking down data silos and using machine learning tools lets agencies get the most of their information.
A major challenge the Defense Department and intelligence community face is managing the massive amounts of structured and unstructured data they possess.
One way to do this is by using artificial intelligence and machine learning tools to help agencies break down their siloed data to use it more efficiently, Eric Putnam, MarkLogic’s executive account lead for the intelligence community, told George I. Seffers, SIGNAL Magazine’s executive editor during a SIGNAL Executive Video Series discussion.
Many contractors like MarkLogic are helping their customers move away from a one-size-fits-all approach to database design that was the hallmark of legacy systems in the DoD and intelligence communities, Putnam explains.
MarkLogic is doing this by “desiloing” customers’ data to bring it into a modernized infrastructure that allows them to make sense of it at the scale they need for the future.
Breaking down information silos is important because there is an explosion of data from sensors and systems that is inundating systems across the government and private sector. This is especially challenging for the DoD and intelligence community because the increasing volume of data threatens to break down their legacy systems, which don’t have the capacity, Putnam said.
To help its customers manage this, MarkLogic is deploying scalable, modernized databases and introduced analytics, artificial intelligence and machine learning tools to help agencies move from legacy “dumb” databases that don’t understand much about the data they hold to a more modern construct where the database acts as a lens, allowing organizations to see what types of data they possess.
This approach has been helpful across the DoD and intelligence community, especially when combined with an architecture called a Smart Data Fabric.
“You want your database to work for you and you want it to be able to answer the questions that you have on top of all the datasets within your holdings,” Putnam said.
Putnam describes a Smart Data Fabric as an architecture approach that allows organizations to access all the data sets across their enterprise. While a traditional siloed architecture might have different databases in a variety of formats, a Smart Data Fabric eliminates this for a more holistic approach supported by AI and machine learning tools.
In a counterterrorism example, where an agency might have information on terrorists, buildings, airplanes and other objects, the system looks for activities across those areas that connect them. This might be a known terrorist being in the vicinity to a certain military facility, which might trigger an alert.
This can’t be done with today’s siloed databases, Putnam said.