Developers, what happens when you design for “system of systems”?
How many of us grasp just how connected we’re becoming? The concept of the “connected world” has enormous – and transformational – implications. Because not only are individual things becoming more connected, each “thing” is of often part of a larger system: a system of systems (SoS).
Narrowing it down, how many of you engineers, software developers and product designers are incorporating SoS-based thinking into your work? If you aren’t, it’s time to start. Because “system of systems” is the new normal.
What is a system of systems?
The simplest definition, is that a system of systems is collection of multiple, independent systems. These come together to form a larger, more complex system. In other words, the sum of the whole is greater than the parts. It’s also a new way of thinking for solving grand challenges. As everything becomes more software driven, how can we integrate it all together?
Connected cars: a system of systems
Understandably, SoS isn’t the most intuitive concept. That’s why experts such as IBM’s Graham Bleakley and Eran Gery showcase it to all kinds of audiences. “When we talk about systems of systems at an event, people know this stuff is out there, but they’re unsure about exactly what it is,” says Bleakley, a solution architect for IBM’s Automotive and Aerospace & Defense divisions. “It engenders a lot of conversations as people try to understand what’s going on, and how they need to think about and approach it.”
The idea is much more complex than just thinking bigger in terms of product development. Consider the automobile. A new, high-end automobile is complex enough with more lines of code than an F-35 Fighter Jet…100 million! Now add in autonomous vehicles and Advanced Driver Assistance Systems (ADAS). These concepts are at the forefront of automotive development, with a high profile in the public eye. How are features like this affecting the development of all vehicles and their connected systems?
Adding concentric circles with autonomous vehicles
Steve Shoaf, marketing manager for IBM’s Watson IoT division, notes that within the automotive industry, automobiles at every level of the market, have always been thought of as complex systems, not just the high-end ones. Every car contains and utilizes other systems, such as drivetrain, suspension, and so on. It’s like a series of concentric circles, with the car itself being the outermost circle.
But autonomous vehicles function within even bigger systems. The first iterations of these are likely to make each car a device within an on-demand transportation system, meaning that sometimes the driver is in control, sometimes they’re not.
That means more virtual circles outside the car. An autonomous vehicle needs to constantly communicate via sensors and the Cloud with other cars, with the manufacturer or dealer, and with roadside data points. “In that sense, the autonomous car becomes a pawn in a much larger game of chess,” Shoaf says. “Because these companies are designing and competing with a system of autonomous cars. This is how these companies are beginning to think.”
Automotive OEMs and other stakeholders are realizing that all this newly available data can be combined with other information. For example, advanced analytics platforms combined with simulation, can determine if there are any faults in the vehicle, or if and how an automaker can enhance a vehicle to improve user satisfaction. This information travels in a continuous loop through the development process. It’s then reflected as changes in requirements or enhancements to existing functionality, implemented via over-the-air software updates. “Today, the automotive OEM’s development process needs to be as interconnected as the vehicles they’re developing,” Bleakley says.
New value chains mean new revenue streams
This new method of development creates a value chain around the cars’ networks similar to smartphones… useless unless they’re part of a network and communicating with each other. As automotive companies start taking a similar approach to the design and development of vehicles as smartphone companies do, the end game becomes much grander than just selling cars.
“For example,” says Shoaf, “if the car participates in an automated parking network, part of the business objective becomes how to maximize revenues from each transaction in which a car might be involved.” Or consider rideshare or delivery services. Transactions are based on trips where the car never parks. “Then, the objective is to monetize the trips consumers take with those cars,” says Shoaf. “In that scenario, the engineering ecosystem has different requirements.”
Obviously, the SoS method is far more complex than established ways of doing business. “Many people think SoS are very simple IoT systems. But you need to use best practices to show that there’s value in doing proper design for these complex systems,” says Eran Gery, a distinguished engineer for IBM’s Watson IoT Customer Solutions. “You can’t just go ahead and throw together IoT systems; the SoS approach is much more disciplined. You need to understand requirements, do proper architecture and analysis, and design for many characteristics. All of this requires significant changes in traditional development cultures.”
Unprecedented complexity will lead to unprecedented cooperation
Of course, no matter how much engineers and product developers evolve their thinking, those pesky consumers — the jaywalkers and unpredictably distracted drivers who aren’t inside the autonomous vehicles — will always be a wild card.
“One of the biggest dangers will be in the transition from the status quo to a more driverless society,” Bleakley notes. “Driverless cars will be more regulated and will continue to better understand human behavior over time. But putting people into an unstructured environment will inevitably create more complexity.”
He believes that solving these problems will require cooperation – perhaps unprecedented – between companies that until now have been fierce competitors. And perfectly comfortable with that arrangement, too. But to realize the full potential of SoS development, they’ll need to agree on standards around safety, back-end systems, and any number of other benchmarks. “It’s certainly a challenge for OEMs, but it also should foster innovation,” Bleakley says. “It will require much more complex testing.”
To this end, he says some automakers are using games such as Grand Theft Auto to accelerate and improve algorithmic testing. “But in the end, it’s about looking at the bigger picture and understanding the interaction between vehicles and the environment,” said Bleakley. “This will improve development and user satisfaction, develop new revenue sources, help us monitor a wider range of things, understand unstructured environments, and then deal with more eventualities.”
To learn more about how Watson IoT is leading this philosophical and technological (r)evolution, join us at the Continuous Engineering (CE) Summit Europe in Munich, May 14-16. You can also visit our continuous engineering landing page