The Office of Intelligence and Analysis at the Homeland Security Department received two deadlines to move to Windows 10.

The first mandate came from DHS Chief Information Officer John Zangardi for OIA to move by the end of September

Dave Bottom, the chief information officer in the Office of Intelligence and Analysis at DHS, said he found a compromise to serve both bosses and at the same time continue his agency’s IT modernization process.

Intelligence Community CIO Sherman launched what he calls the second epoch of IC IT Enterprise (ICITE) earlier this year that includes a more federated approach to adopting common services.“Part of that then gets to adopting shared services. As you know, the intelligence community has signed an agreement with Microsoft for Azure and Office 365 within the top secret environment. We plan to migrate to that and adopt those shared services,” Bottom said on Ask the CIO. “Then we will continue the adoption of shared service from ICITE that started before I got in this chair, but really was one of the reasons I was asked to come to DHS to accelerate that. We are adopting commercial cloud services, that the CIA is the executive agent for, and the IC GovCloud, which another IC partner is responsible for.”

The CIA is using the commercial cloud offering from Amazon Web Services, while the National Security Agency is providing the GovCloud.

For Windows 10, Bottom said OIA is taking the unclassified image and doing some tailoring on the desktop and the back-end infrastructure to make it work in the top secret environment.

Bottom said since OIA has a foot in both the intelligence community and the civilian agency worlds he is finding a way to balance the requirements while ensuring his IT infrastructure and mission needs are improving.

That is why OIA is leaning heavily on the cloud and shared services whether it’s for top secret requirements or for back-end administrative needs.

“One of our goals and objectives is to make sure we have those services instrumented in a way where if we do have an issue, we can do the right amount of troubleshooting and fault isolation to get service up quickly,” Bottom said. “It’s much harder to do that in a shared services environment than if all the services are within your own data center. That is our focus area to make sure we are doing the right instrumentation as we adopt those shared services.”

Bottom said the move to cloud and shared services also requires OIA to do some business processes match the technology and not the other way around as well as address their approach to budgeting.

The Office of Analysis and Intelligence manages a working capital fund to provide common services for other DHS components. But instead of buying hardware or software, the move to cloud and shared services is more about buying time and dealing with spikes and troughs.

“How can we be more agile in the managing of dollars that you might not know how they will be spent the next month so changes to business processes is a huge play here,” Bottom said. “In large part, this is cost avoidance and changing your cost profile. You are getting more for your dollars, especially with capabilities of Microsoft and AWS through commercial cloud.”

At the same time, the move to cloud and shared services is helping the Office of Intelligence and Analysis more quickly adopt artificial intelligence and machine learning.

Bottom said through these emerging technologies, analysts and collectors can better “tease out” patterns or behaviors that could be potential or real threats to the homeland.

“Without the services the IC is providing under ICITE, that would be really cost prohibitive and take us an extremely long time to do,” he said. “We are in a position now where we can adopt those services quickly, deploy the tools and get the data in there to enable our analysts to tease out the patterns that require deeper insights.”

Bottom said the biggest difference around AI and machine learning is the use of models to capture an analyst’s knowledge and experience and apply machine speed to the data.

“Without that type of capability, it really will be hard for our analysts and operators to focus on the most urgent or critical thing at that particular point in time,” he said. “We are spending an awful lot of time on the data side. We still have a lot of data in their own siloes and if you are going to do effective machine learning, I haven’t found an effective machine learning algorithm that doesn’t require the data to be in at least one physical or logical place. So how do we strike the balance between the costs and time associated with moving data from one silo to a centralized place where it can be compared is an area of significant interest and investment for us.”

Article via FederalNewsRadio.com