The Intelligence Community (IC) and the Department of Defense (DoD) leverage cloud architectures to deliver data to missions faster and securely while creating efficiencies. The IC embraced the cloud early through Commercial Cloud Services (C2S) programs; now, it matures from a mix of on-prem and cloud with a single cloud provider to a hybrid and multi-cloud architecture under the Commercial Cloud Enterprise (C2E). And while this brings benefits, it also brings a few challenges.
On Feb 17, the USGIF GEOConnect Series Main Stage hosted an all-star panel of cloud computing and GEOINT experts led by moderator Rebecca McBride, Director of Strategic Partnerships, Technology Transformation Services, GSA. They discussed how GEOINT users and providers are maturing cloud technologies while examining the opportunities and challenges with moving into hybrid cloud and multi-cloud environments.
Full GEOConnect Series Main Stage Session. To view the full session, you must fill out the login form.
Capturing the Value
From enabling mobility to realizing enterprise-scale transformation, cloud technologies empower GEOINT professionals to solve the most challenging problems and deliver intelligence to customers when, where, and how they need it.
A lot of the focus to date on cloud technologies has been on lift-and-shift migration. But according to Larry Socher, Accenture Federal Services Cloud, Infrastructure, and Edge Lead, the real value of cloud is the innovation and investment of hyperscalers in middleware and platform as a service (“PAAS”) functions. It’s about how organizations modernize applications and data to take advantage of these innovations. That is really where the power is, according to Socher. “While there are definitely benefits of leveraging the hyperscale and [infrastructure as a service] for doing some lift and shift migration to the cloud, it really needs to be a part of a broader app modernization strategy,” he said.
Socher noted that many commercial clients migrated quickly to the cloud but never obtained the full value out of it—the modernization of the applications to take full advantage of that innovation cycle.“It’s about how you modernize your applications and data to take advantage of hybrid and multi-cloud environments,” Socher said. The most successful early adopters of cloud did not only rely on a strategy and expert advice—they went on to extract value from hybrid and multi-cloud experiences, he added. Additionally, it is also important to note that one challenge is whether organizations should choose one hyperscaler or multiple. According to Socher, in some cases it’s better to bet on one hyperscaler for various reasons, such as the dilution of skills that is required to work with different cloud environments, the amount of work necessary to operationalize the environment for different providers, and the challenges with data gravity when moving data between different cloud environments.
In this regard, the IC can learn lessons from industry about the kinds of factors to consider as the community accelerates into this next important phase of the cloud journey. While this next phase will be challenging and complex, the IC has a way forward, and can learn from commercial organizations who tried and succeeded in cloud adoption–as well as from those who tried and faltered in their cloud efforts.
The National Reconnaissance Office (NRO) has been migrating their workloads and taking advantage of cloud technology. According to Col Erich Hernandez-Baquero, Principal Deputy Director, Ground Enterprise Division, NRO, as the organization began moving mission systems, cloud performance was critical; this was particularly important for the resiliency and reliability of the systems. However, one challenge the organization faced was figuring out how to harmonize their systems to support their users. According to Col Hernandez-Baquero, their hybrid architectures have different Operations and Maintenance (O&M) and sustainment models.
“We may have an on-prem platform provided by a single contractor who hones the full stack — from the network to the actual application. Monitoring this performance would be different than monitoring software capabilities deployed to the cloud, where different layers are now involved,” said Col Hernandez-Baquero.
Additionally, some challenges also arise when dealing with cost. Unlike on-prem systems, it’s challenging to forecast price and predict performance on cloud architectures. According to COL Hernandez-Baquero, NRO’s challenge is to get accurate forecasts to budget for the cost.
NRO has completed a significant amount of work to bring those elements together and provide a unified front for their operators when an anomaly occurs or a fault happens. But that’s not trivial; it takes a lot of work, and the NRO has continued to improve in this area.
According to Jason Weiss, Director of Software Modernization, DoD CIO, it all starts with agility. When you think about an agile team, you’re looking at those capabilities you want to deliver, not forecasting the system’s state in the future.
“We used DevSecOps as an approach that we were aspiring to move to very rapidly, but even in that scenario, DevSecOps is not looking at the future. We have to differentiate between the solution architect on the DevSecOps team and the enterprise architect looking at the big picture,” said Weiss. “In my opinion, a lot of enterprise architects are viewed only from a technical lens and what their technical skillset is. Realistically, enterprise architects have to have heavy business acumen. They have to understand the mission, and they have to understand how a contract vehicle is going to accelerate that mission.”
Increasingly, the ultimate challenge is optimizing more complex environments. So, it’s one thing dealing with static virtual servers, but as we start to adopt more Cloud Native architectures, more distributed microservices, we get more dynamic and ephemeral environments.
Therefore, according to Socher, managing cost and performance in a hybrid environment revolve around two significant complexities.
The first is if you look across hybrid, it’s easy to see public costs, but it’s hard to get a good handle on what private or data centers cost. So, do you have tools in place that provide a cost basis for what it takes to run your own data centers or private clouds?
Secondly, ask yourself how you simultaneously optimize service levels, resiliency, availability, performance, response times, and cost. Ultimately, according to Socher, humans just aren’t going to be able to do that optimization.
“We have to use artificial intelligence, machine learning to do that in the long run,” said Socher.
Building a Framework
In the past year, the USGIF NRO Industry Advisory Working Group began to explore the hybrid cloud and what it signifies to the GEOINT community. Chris Arroyo, Director for National Security Programs, CloudBees, was part of the working group’s hybrid cloud action team. The team built a framework for government to use when approaching the hybrid cloud.
Arroyo and his colleagues compiled a list of questions to be used as a benchmark so that each program can figure out what makes sense to migrate based on cost and bandwidth. Secondly, when discussing migrating applications into the cloud, Arroyo noted, it is important to audit those apps and find out whether they are legacy or cloud-native apps to determine the kind of lift needed to refactor the application for the cloud.
Finally, when considering hybrid- or multi-cloud scenarios, there is the issue of portability — what happens when you want to exfiltrate data and the related applications out of the cloud back to your on-prem data center or move it to another cloud? What does that cost? What is that time requirement? Is it doable? Having all of this information upfront will help you decide if and how you migrate capabilities to the cloud, Arroyo said.