Unravelling the Secrets of Simulation as a Service

Many would argue that the concept of Simulation as a Service is an enigma. Even as more sectors of international economies embrace this nascent technology, so little of its workings are actually known. To shed more light on this elusive subject, we had a chat with Wim Huiskamp, a Scientific Advisor to the NATO Modelling and Simulation Group (NMSG).

Following a 37-year stint at The Netherlands Organisation for Applied Scientific Research (TNO), where he serves as Chief Scientist of Modelling & Simulation in the Defence Research Branch, Wim is well-versed in the data-oriented worlds of modelling and simulation. With a professional focus on M&S architectures, he has made strides in developing interoperability standards and also introducing better modelling concepts, both in his native field of defence and other related domains.

What is simulation as a service?

At its simplest, Simulation as a Service is a cloud-based service used for decision-making and training. It's the ability to rapidly create and optimize virtual physical environments using simulation models. These models can often be altered at will, with interfaces that allow users to see how application models affect each other.

Creating models is a challenge in itself: "By observing the behaviour of real people in real scenarios, we can gain insight into their drivers and mental processes," Wim explains. "The data from these observations is captured and translated into models that essentially map out behavioural guidelines and repetitive patterns which are then used in running simulations."

But the scope of Wim's work goes beyond that.

"At the NATO M&S Group, we develop the architecture that makes simulations work. We develop standards for Metadata that describe these services. These standards tell you everything about a service. From what it can do to the extent of modification it can take, it's essentially where all search queries for services are cross-referenced. We also develop standards for achieving interoperability between services."

The benefits of Simulation as a Service

Versatility is one of many perks Simulation as a Service offers. Wim likens simulations to reality - reasonably constant. But even reality with all its laws and constants is wildly fickle. Simulation as a service accounts for this unpredictability by making provisions for swift adaptability to ever-changing real-world factors.

"Systems change. Mission areas change. Opponent behaviours change. And when they do, we need to be able to quickly add features, or modify our simulations to reflect that change. The innate flexibility of simulations makes that easy and allows us to be more dynamic and responsive in meeting the needs of its users, however sudden it might be."

The technology also places heavy emphasis on efficiency. As more defence programs turn to simulated environments to train their personnel, new services don't need to be built from the ground up.

"The need for reuse of simulated services has always been existent. Luckily, this is one void that's filled by Simulation as a Service. Rather than reinvent the wheel or build new simulations from scratch, it's possible to reuse the service already developed by someone else."

Its pros hardly end there. From having no harmful impact on the environment to drastically reducing cost margins, Simulation as a Service offers training programs and overhauls that were once that of wishful thinking.

"The added benefits of running services in a cloud environment are endless. For one thing, you end up with a smaller local footprint. Plus, you incur fewer maintenance costs. There's also the eliminated need for buying hardware. Not to mention, since weather, equipment, or even personnel aren't deciding factors anymore, training is more flexible too."

Challenges to widespread adoption

However, the path to a future run on Simulation as a Service isn't without its obstacles.

Discoverability and scalability are two of the several technical issues the technology faces. With so many cogs to the Simulation as a Service wheel, coordinating component solutions to create (compose) the desired simulated environment has proven difficult.

"Because all services must work harmoniously, we often have to build solutions that make sure they do," says Wim. "And that could have different meanings in different scenarios. Sometimes, the simulations demand that we start more services and scale up. Other times, it's the opposite. We do our best to manually manage the situation but it's certainly not a solved problem to fully automate this process."

Security is another challenge the service based approach faces. With the majority of simulations being hosted in data and compute clouds that are usually miles away, distrust tends to rise in many users - especially in the defence and security domain.

"When simulation services are run on a Microsoft or Amazon cloud, there's a good chance their servers will be in another nation. One you probably won't know. Because this confidential data is in a cloud they can't geographically pinpoint, people tend to get nervous."

But neither of these problems are insurmountable. Wim shares how the use of hybrid solutions is rising in popularity. These solutions allow sensitive data to be stored on private local servers while the unclassified parts are sent off to commercial clouds. And while that's perfectly viable, the NATO Scientific Advisor recognises that these novel solutions may require a change in mindset for organisations.

"It's a totally different business model. A cultural change, even. We all are used to buying services and running them on our own computers or in our own buildings. We are accustomed to being in full control. Going from that to entrusting your data to a third party can be a daunting transition. But it's one that companies must make nonetheless to gain new capabilities."

Architecture for distributed simulations

As global organisations extend their branches into even more countries, the need for training simulations is at an all-time high. But with international branches in different time zones and sometimes, different continents, how can these simulations be coordinated according to a common standard?

NMSG's work on Specific Reference Architectures comes to the rescue.

"Distributed training in a multinational context demands more than just exchanging information during an exercise. It involves sharing the scenarios, environments, terrain databases, and all the assets you would find in this synthetic world.

"Without these validated interoperability standards, distributed simulations would be an uphill battle. Basic yet vital actions, like seeing a synthetic footprint left in simulated mud by a player training from the other side of the world, would be impossible."

Cyber effects in a distributed synthetic environment

Regardless, even with thorough frameworks like the High Level Architecture (HLA), interoperability standards are far from complete. The evolution of real-world threats has spurred an equal change in the type of simulations trainees undergo. The event of cyber attacks is one such example.

"What does the victim see? What do they notice? How do we represent that effect in the simulated world? These are the questions we strive to answer when dealing with cyber effects in distributed synthetic environments."

As we progress further into modern times, training and simulation innovation programs like Wim's help its users stay ahead of the curve. Whether it's through crisis management, decision making, or behaviour modelling, organisations have turned out to be better for it.

To learn how we can design your custom training and simulation solution, get in touch today, or fill out the form below.