Clouds on the S&T Horizon

Group Editor Marty Kauchak provides an overview of the US Defense Department’s quickening pace of developing cloud-enabled training and education capabilities.

The US Defense Department is doing nothing less than disrupting current concepts of anytime, anywhere training delivery – by beginning a transition to cloud computing. Taking their cue from advancements in cloud computing observed in other DoD enterprises and the commercial sector, training organizations are seeking to use cloud computing to deliver diverse computing services (servers, storage, databases, networking, software, analytics, others) over the internet (“the cloud”). In other words, the heavy lifting is done in the cloud, not by the client. Early US military activities in this space viewed by MS&T provide the promise of more innovation, economies of scale and other returns on investment for servicemen and -women.

Overcoming Hardware Limitations

SAIC is one of four simulation and training sector companies surveyed by MS&T which are bringing different competencies to bear to allow prospective military customers to migrate S&T to the cloud.

The military sector’s demand for cloud-delivered products and strategies is driven by its requirement to provide training to the learner at the time and place where they need it, often beyond the traditional classroom, according to Jeffrey A. Raver, SAIC vice president for training. After acknowledging that distance learning, web-based simulations and other current strategies enable “anytime, anywhere training,” Raver responded, “Those technologies can provide distance learning, but they are often constrained by end-user hardware limitations. Cloud-delivered training solutions overcome those limitations, delivering everything from simple interactive multimedia instruction through high-fidelity simulations to any device with a network connection.”

SAIC’s Integrated Training Edge (SITE) internal research and development efforts have yielded several cloud-based simulation projects. One cloud-hosted project is the Synthetic Cyber Sandbox, which garnered much attention from delegates at I/ITSEC 2018. This project originated to support internal SAIC distributed workforce development and collective cyber training as well as Simulation as a Service (SIMaaS) initiatives.

Raver furnished some early returns on investment from SAIC’s use of this internal cloud-based simulation project. “Centrally managing software, databases and user access is more efficient and effective, especially overcoming cybersecurity concerns and more easily collecting student performance data. Additionally, cost savings are realized due to reduced equipment, manpower, and sustainment costs.” Further, the cloud-based architecture permits access to the same content by a larger learning audience, when compared to legacy delivery methods.

Brian Sieck, the Synthetic Cyber Sandbox product manager, noted that a common topic among prospective customers is the learning content migration process of integrating a current simulation or like content into the cloud. “While the complexity of the simulation governs the resources needed to complete the migration, security requirements, security authorization, and governance are often drivers of the timeline,” he added, and estimated two to six months to complete that process, depending on these factors.

Reflecting a development MS&T sister publication Civil Aviation Training (www.civilaviation.training) has observed in the adjacent civil aviation sector, SAIC’s Sieck pointed out the importance of
data analytics to support cloud-based learning.

This capability would permit instructors at a central location to finish evaluations, tailor the training to learners and complete other requirements for larger numbers of widely dispersed students. “All of that information is collected in a data repository and can be analyzed more efficiently, and ported back out on a demand basis from the instructor,” he emphasized.

Cloud content enables this distributed training." Don Bray, Raytheon Intelligence, Information and Services Cyber  training director. Image credit: Raytheon.

Raytheon Intelligence, Information and Services is also stepping up its presence in the cloud-enabled learning space, with activities which included the unveiling of its Persistent Cyber Training Environment (PCTE) demonstration at I/ITSEC in December.

Don Bray, the director of cyber training at the Raytheon division, observed his office is increasingly seeing requirements from the government for distributed training solutions which require cloud-based technologies. The Persistent Cyber Training Environment, or PCTE, for the US Army is one example. “PCTE will use virtual machines connecting to different cyber ranges to create a cyber warrior training environment. Raytheon developed a geographically distributed PCTE solution located at our facilities in Orlando, Florida and Dulles, Virginia but available anywhere,” the subject matter expert explained.

Raytheon’s Bray highlighted the business case for a military customer to migrate content to the cloud, noting that moving this content to the cloud will enable training from anywhere. “Traditional training limits where and when our customers can learn,” Bray emphasized, adding, “But, specialized training is often needed at forward-deployed locations and on an ad hoc basis. Sailors deployed aboard a ship or forward base, for example, may need refresher training for operating a new radar or new software. Cloud content enables this distributed training.”

Raytheon applied corporate competencies to help its customers migrate to the cloud, taking advantage of its expertise in automation. “Our suite of tools automates managing the cloud environment, allowing us to skip ‘old fashioned’ server administration tasks to focus on delivering code,” he concluded.

Bohemia Interactive Simulation's cloud technology suite enables combined arms and multi-domain scenarios. Image credit: Bohemia Interactive Simulations.

Simulation Data on Demand

Bohemia Interactive Simulations (BISim) is on contract with the US Army to prototype technologies that will allow the service to deliver training to the point of need, wherever that is around the globe, via the cloud. The service’s expectations for training from the cloud, in one instance, are noted in the statement of objectives for the service’s Synthetic Training Environment (STE) program, according to Pete Morrison, the company’s Co-CEO. “In a similar manner to Bing or Google Maps, which ‘magically’ connect you to cloud-based servers and provide you with a service, the US Army wants simulation to be that service,” Morrison explained.

Moving delivery of training to the cloud will represent a huge course change in military learning strategies. Morrison acknowledged it will be “quite massive” with significant opportunities to increase the effectiveness and efficiency of simulation. “As a whole, the US Army is not collecting data on the performance of soldiers across the service in simulation. It is also not centralizing terrain in a repository that can be accessed at run time by all simulators. Now SE Core [Synthetic Environment Core] develops terrain and delivers that content sometimes by hand, other times by the internet/wide area network, to the point of need,” Morrison pointed out, and added, “The cloud will enable the simulators at the point of need to simply connect and pull down that terrain data on demand.”

Projected returns on investment in investing in cloud-based delivery for simulation users? Morrison noted it currently takes up to six months to request terrain through SE Core. “The cloud will permit us to accelerate that – there will be a persistent, virtual Earth that is being updated continuously that your simulators will connect to and pull-down current data that is needed immediately. In-theater service members will be updating that database with what they are seeing in the field. Others at home station can train on that the next day. This is, I would say, one of the biggest advantages of cloud-enabling simulation.”

The main technical impediments to the Army and even other military and commercial customers quickly migrating instruction to cloud-based delivery are bandwidth and cyber security. “While both bandwidth and cyber security are certainly issues for the US Army,” the BISim leader concluded, “I have never seen as much momentum to achieve early implementation in the field as I have seen on STE. The Army leadership has the vision and the will to initiate significant change. I believe there is a revolution coming, and coming soon, as the Army transitions away from ‘heavy clients’ in battle simulation centers to cloud-based simulation with ‘thin clients’ and massive scale.”

Monitoring and Motivating Learners

Beyond terrain databases and similar content, the US military customer is also looking to move training services to the cloud. In one instance, training organizations are revisiting their conventional architectures for learning systems and realizing their tutoring systems “are not monoliths. They consist of discrete components that work together, which supports a cloud-based services way of thinking about training,” observed Dr. Benjamin Bell, president of Eduworks.

The Corvallis, Oregon-based company is responding to the expanding expectations of the military customer in this sector, beginning to build for the US Air Force a “detector that can flag” when the learner is becoming less motivated and disengaged. Bell observed that, in the current mindset, such a capability would be built and inserted into the tutor, to flag disengagement and make recommendations to recover the lost motivation. “When you think about moving that into the cloud, you realize when you build that kind of appliance to detect lapses of motivation, it does not need to be joined with a specific tutoring system. It can be a generalized appliance,” Bell added. Eduworks’ vision for this effort “is to create a cloud-based microservice that any learning environment can interact with through an application programming interface, as a service that can say to that learning environment: Hey, I think you are ‘losing’ your airman, he or she is disengaged and here are some generic recommendations to regain that level of engagement.”

In the initial proof of concept work demonstrated for the Air Force Research Laboratory, the core capability to detect lapses of motivation was integrated as an independent and interoperable entity in two different learning environments. Further work with this project started at AFWERX last fall, with the expectation the proof of concept would be completed this January.

Another cloud-based service garnering attention at Eduworks is CASS (Competency And Skills System). “We’re really excited about this because of the transformative impact it can have on the US DoD and human capital management sectors. It permits organizations to manage workforces in terms of competencies,” Bell added. One application of CASS would support a service member transitioning to the civilian workforce after accumulating a number of credentials, by permitting a civilian organization to better understand the civilian equivalent competencies represented by the diverse military credentials. An important enabler of CASS is artificial intelligence (AI), which “auto aligns” the competencies, permitting algorithms to parse through the content and make recommendations on the degrees and other credentials present.

The Eduworks leader provided a view of the transition of content to a cloud-based construct: “Widespread adoption and scalability will come when we tolerate incremental progress in the sophistication of the training. This is a big leap to solve the very complicated, universal tutoring problem, and at the same time move it to the cloud. If we appreciate incremental, lighter, pedagogical solutions, we can address scalability and broad adoption.”

SAIC's cloud-enabled Integrated Training Edge (SITE) provides flexibility to safety take learning beyond the classroom. Image credit: SAIC.

AI Teachers and Megacities in the Cloud

SAIC views the continued, concurrent development of its artificial intelligence portfolio as important to maturing its cloud-delivery strategy.

“Unlike classroom training, an instructor may not always be available for cloud-delivered content,” Raver said. “AI provides greater efficiency by allowing the students to interact whenever and wherever they want, without having to rely on the availability of a live instructor. An AI ‘teacher’ provides students with feedback and immediate response to the things they are learning, similar to what would normally be happening in a classroom.”

Another critical, emerging underpinning of cloud-delivered training is cloud scalability – aggregating computer servers to do significant amounts of processing. Morrison said BISim will be using cloud scalability to simulate millions of high-fidelity entities in a “mega-city.” “Until now that has not been possible” because contemporary computing processes cannot handle the scale and hence entities are typically aggregated, i.e. the simulation “cheats” by simulating groups of entities as one. Morrison revealed, “We’ll be demonstrating this in May to the US Army, with 2-to-3 million entities, with complex behaviors. And you will be able to have a scenario happening within that broader context – so your human-in-the-loop squad or platoon can be operating in a living, breathing mega-city.”

Confluence of Activities

The US military-industry S&T team is accelerating its pace of activities to migrate courses, simulations, services and other learning-related content to the cloud. Along this journey, the team is casting a wide net to bring to bear automation, AI, data analytics and other capabilities to elevate service learners’ activities beyond the contemporary notion of anytime, anywhere learning. Initial projects have yielded the promise of returns on investment for the Pentagon. MS&T will continue to monitor and report on activities in this expanding sector through the magazine pages, the militarysimulation.training website and e-newsletters.

Originally published in Issue 1, 2019 of MS&T Magazine.