Data integration in government is not easily achieved given exponential data growth, inherent organizational silos within agencies and unique mission needs. Dave Rost, a senior account executive and Defense Department subject matter expert at MarkLogic, said the individual silos have their own way of doing business and don’t integrate across the entire agency.
“What happens is you can never get a full single source of truth from your data, and when you try to integrate that data, it takes sometimes years to complete,” he said. “There are so many different types of file stores and databases, from relational to graph [databases] to individual share drives that people store their data on. Trying to get that information all together to find one comprehensive look at that data sometimes takes years to do and by the time you get that answer, it’s either outdated or the mission has changed to where you need to do something different.”
The best way to solve this challenge, according to Rost, is to incorporate an enterprise data layer where agencies can have a single view of all of their information. This single view enables agencies to bring all their data silos together so they can quickly run queries and realize immediate benefits from all their information sources.
Rost said MarkLogic’s enterprise data layer, which runs on Intel technology, pulls agency data together quickly and then the agency can build a variety of applications on top so the information can be viewed in different ways.
“The benefit to that is having one data layer that serves multiple applications, so you can have all that information fused together at the data level, and then each silo or each agency or department of that agency can have their own view and use that same data,” Rost said.
So, how should agencies approach a move toward enterprise data management?
Rost cited one of MarkLogic’s customers to explain the path.
“One approach that one of our customers, [Defense Department’s] Unified Combatant Command, took was to look at the data they currently have and then provide a mechanism to adapt that in the future as new data comes online or new things happen,” Rost said. “And they stood up a MarkLogic data layer and they started ingesting different pieces of data – intelligence data, mission sets data, HR data, contracts data – and they put that into a single view. When they did that, they were able to find additional value in that data – things that were relating across those different data sets that were never looked at together at the same time before – and they were able to do that because they started fusing that data from the silos into the single source view.”
The impact this can have on an agency, particularly from a human capital perspective, can be significant as agencies regularly task workers with running data searches and compiling findings.
“Getting that data together and finding that answer used to take about 80 percent of the time just trying to find the data, and then they have only about 20 percent of the time left for either analysis, curating the answer or creating the final product,” Rost said. “Now we can flip the paradigm to a ratio of 20 percent getting the data and 80 percent creating the product or analysis, which is greatly impacting the mission speed and the rate at which we can operate. Additionally, it’s taking the human capital cost of being able to produce quality products and improving that greatly.”
Rost also said MarkLogic and Intel are ensuring data agility and flexibility so government decision-makers can use and access all the information their agencies have gathered – regardless of how or in what form it was collected.
“Since MarkLogic is a flexible database platform, we can load that data as-is,” he said. “A lot of agencies when they start trying to take this challenge on of putting an enterprise data layer together, they try to get that perfect solution and they have to model that data just perfectly and spend years doing that. And when they do that, they can’t change it. They have to start all over again as a new data source or a new mission set comes up. With MarkLogic you ingest it as-is. We don’t care what type of data it is. You don’t need a model of risk diagnostics. You put it in the databases and immediately get results from that.
“One of the hardest things is making the data available. Data in silos becomes locked in almost. Trying to expose that data and getting it into anything else, regardless of the end solution you choose, it’s always a difficult task,” he said. “With MarkLogic, you don’t need a model for the data, we can just take it. It doesn’t matter. You can leave it in there if you want to and we can provide reach back into it, almost like a federated search or reach-back capability, and that removes that risk from them. It gives them the comfort level where they can invest in MarkLogic and immediately see results from it.”
Rost pointed to one agency customer that was able to realize immediate benefits with this approach.
“One of the agencies that I think is really applicable for that is a national level intelligence agency in which most of their data was residing in about 30 databases and users would have to go query each database to make sure they had a complete picture of it,” he said. “They started bringing all that data together within MarkLogic and now they have one source where they go to it, one piece of information they get that complete answer immediately.
“We were literally six weeks from project start to deployment and then we took another six weeks later on where we moved it from premise to cloud – because we completely run on the cloud – and putting the data in the cloud allows not just cost savings but opens up the door to what you can do in the future.”
Rost also acknowledged the potential for error when running queries on dozens of databases and the impact that can have.
“One of the problems you have with those different databases is no search language is the same,” he said. “Each database is going to have a different type of search, different type of results coming back and you have to factor in human error where you misspell words and if you misspell a word in the search you’re going to get different results back. You may be missing a key piece of information to fuse together. With MarkLogic, if you misspell a word, we automatically understand that in the database and we are able to give you all the search results. Take the example of ‘vehicle,’ which is a commonly misspelled word. In a lot of reports, it’s a simple word. But since it’s so commonly misspelled, if you’re not searching for all the misspellings, you could get that result wrong. MarkLogic’s solution can account for those different types of misspellings with the way we do our search.”
Rost said agencies working toward integrating their data don’t need to have the perfect solution upfront – they just need to get started.
“Start with something, build on that and reiterate it,” he said. “With MarkLogic you don’t need the perfect model like you do with everywhere else. Get the data in there, start finding the value and each separate system can then build their application on top of it. But you immediately start seeing value, instead of creating this giant monolithic plan that takes three or four years to do. It’s going to change in three or four years, it always will. We change almost daily now. And if you don’t have that flexible way of doing business, then you’re just going to fail, and it’s not going to be cost effective for you.”
Jason Fornicola joined WTOP and Federal News Radio in February of 2014 as the Director of Custom Media, where he manages the stations’ sponsored and custom content initiatives. Fornicola provides clients with access to high-quality content consistent with the stations’ standard of excellence and works to solve clients’ needs through a multi-platform approach which includes on-air, digital, video and social media. His journalism and social media experience support the stations’ efforts to ensure the client message is conveyed in a way that will achieve their marketing campaign objectives.
Dave Rost, Account Executive and Subject Matter Expert on Insider Threat, MarkLogic
Mr. David Rost is the Department of Defense (DOD), Intelligence Community Manager responsible for the all MarkLogic Military Intelligence and DIA Programs. Mr. Rost oversees the area growth, maintenance employing the MarkLogic capabilities into the DOD/IC business and intelligence processes. Mr. Rost is responsible for client management, operational leadership for critical mission areas and financial oversight.
Prior to MarkLogic, Mr. Rost worked as the Senior Program Manager for the Strategy and Business Development team under the Boeing Integrated Information Systems Business Unit. Mr. Rost also served as the Senior Director for the Insider Threat, Cyber and Counter – Weapons of Mass Destruction (C-WMD) Operations. As the Senior Director and Program Manager, Mr. Rost oversaw and provided support for the TAC Program to a variety of Department of Defense (DoD) organizations, employing the TAC capability into their business and intelligence processes. Mr. Rost was responsible for the development and implementation of Boeings enterprise Insider Threat data layer and system. Mr. Rost held numerous other positions in his tenure at Boeing providing support to the National Joint Operations and Intelligence Center at the Pentagon, the Defense Threat Reduction Agency, Army Cyber Command and various Special Operations Organizations. This support has ranged from introduction and advanced training of the TAC system to business and intelligence workflow process assistance and providing data and information management consolation
Prior to joining Boeing Defense, Space & Security, Mr. Rost worked as a contractor for the Defense Intelligence Agency (DIA) supporting the Afghanistan mission at the office for Asia Pacific Analysis. During that tenure Mr. Rost worked on a project requested by the President of the United States that specifically assisted in measuring the stability of Afghanistan – directly affecting U.S. Government operations and fiscal planning in country. Additionally, Mr. Rost and a colleague obtained funding and initiated the first ever DIA dedicated social network analysis team to investigate and visually depict critical human networks of interest to the Afghanistan mission. This program was widely successful in providing critical node analysis to operators.
Mr. Rost served in the United States Army and New York National Guard for over twelve years as an All-Source Intelligence Analyst. While on active duty, Mr. Rost deployed several times including an 18 month rotation into Iraq where he served as the Non-Commissioned Officer in Charge for the Intelligence section at the Division Tactical Operations Center. In this capacity Mr. Rost coordinated division intelligence assets and dissemination of information for current operations and future operational planning.
Mr. Rost holds a Masters in Business Administration from the University of Maryland, Bachelors of Arts in Intelligence Studies, and a Bachelors of Science in Criminal Justice.