By Alan Levine, Vice President - Architecture, Strategy and Product Management
Sean Parham, Vice President of Internet of Things Systems and Software at Cisco Systems
Brian Dummann, Chief Architect at Emerson Electric
Part of the Oracle Experiences in Enterprise Architecture article series
One of the biggest developments shaking up the business world is the Internet of Things (IoT). For example, in the manufacturing and industrial automation sectors, many objects are now equipped with advanced sensors that can send and receive data, leading to a tremendous variety of information and vast new data sets to learn from. What role do enterprise architects play in helping organizations create applications that can utilize this data? What are the specific problems that architects should focus on?
To answer these questions, attendees at the 2014 Oracle Enterprise Architecture Summit heard from two experts in the IoT field: Sean Parham of Cisco Systems and Brian Dummann of Emerson Electric. Their respective roles differ but they had some significant insights in common.
Sean Parham leads new ventures in the IoT Group at Cisco. He and his team are focused on the creation of new products for the emerging IoT market. He leads product and business strategy, engineering, business development, service operations, and product management. As he pointed out, so far there are only rudimentary standards in the brave new world of industrial IoT. He sees parallels with the early days of the Internet when there were many conflicting network protocols, from TCP/IP to DECnet to AppleTalk. "It took a long time to sort this out and establish consistency so that computers and terminals could easily exchange messages and data," he said.
This is the same problem facing today’s IoT community as IT professionals focus on standards that shape their businesses. "At Cisco, we currently focus on nine different IoT verticals, each of which has its own sets of standards and modes of communication with its unique end customers," he added.
Another challenge facing enterprise architects is the sheer volume, velocity, and variety of data. IoT entails a major shift in scale—from enterprise applications that might have a few hundred endpoints to industrial applications with potentially millions of endpoints.
There are many cultural challenges as well. Industrial engineers and line of business (LoB) leaders tend to think too narrowly about their products. Enterprise architects must help them consider the larger business processes in the context of connectivity, security, availability, analytics, and integration with enterprise applications.
Other complications arise from the very different types of users that are impacted by industrial IoT, which may be much different from what IT professionals have encountered before. For example, a factory worker who is accustomed to a familiar human machine interface on a computer terminal might suddenly be interacting with data from dozens of machines and hundreds of sensors on an assembly line. This causes a fundamental shift in the way enterprise applications interact with that worker including different types of conductivity, different types of security, different types of data, and new constraints governing innovation and safety.
"It’s not always people peering through windows into our applications," Parham explained. "It is more about how IT can become pervasive in the overall operation of the things that people do. It often involves a completely different kind of transaction model and a completely different kind of audience."
The end points might be disk brakes on a bus or electric transformers in a smart grid. The interface points might be equally unconventional, such as a specialized diagnostic device in an automobile shop or a ruggedized terminal at the top of utility pole. "Basic IT assumptions about displaying and interacting with the data no longer apply," Parham added. "Today’s application-enablement platforms must give us different ways of looking at the infrastructure and the people who interact with that infrastructure."
In summary, the world of the industrial designer is much different from the world of traditional IT. Information needs to be captured at multiple places and end points, and then consumed by new types of equipment and users. To design effective solutions in this new world requires different state models, different power, and an appreciation for the different bandwidth, latency, and security considerations.
Brian Dummann, Chief Architect, and his colleagues at Emerson are wrestling with architecture challenges that arise when different business units deploy similar technology simultaneously. Emerson serves several key markets, from the oil and gas industries to the data center and industrial automation markets, all of which are embedding sensors in their products to acquire real-time data. However, in many cases they are embarking on these ventures independently of one another. Emerson’s enterprise architects are helping these business units move forward in a cohesive way so that they learn from one another and so that engineering and IT departments do not have to deal with unnecessary diversity in implementations.
For example, the architecture team at Emerson is working to encourage common standards across five major business groups and more than 50 business units that are responsible for creating the company’s intelligent products. In addition, the enterprise architects are helping to define key components and frameworks. This includes promoting a common view of what the Internet of Things means across the company, such as how to structure machine-to-machine interfaces, build messaging into its products, and connect various types of devices so that those devices can send, receive, cache, and forward information.
These architects are helping business leaders at Emerson gain a broader view of IoT solutions with respect to information flows, data lifecycle challenges, software development roles, and related issues. They have constructed a capability framework to ensure that each business unit is "speaking the same language" about new capabilities, correctly identifying and leveraging opportunities, and solving common challenges together. "Models are great, and architecture can do a great job in bringing together a common set of terms and definitions," Dummann said.
As enterprise architects look across this diverse IoT landscape, their task is to establish frameworks and reference architectures to help manage complex activities such as sensor-based data collection, high speed data transfer, complex event processing, data security, cost effective storage, and real time analytics. They must help are help their organizations—and the industry at large—by developing foundational architectures and strategies to help companies bring IoT solutions to market as well as by ensuring adherence to and governance of those architectures.
Organizations that put the necessary effort into understanding these issues derive tremendous value. However, IT professionals need to frame the discussion differently from how they framed it in the past. "They aren’t always talking with other technologists, like they were when they were sorting out network protocols," Parham noted. "They need to understand how people in an IoT workflow behave and then create interfaces that they can use," whether it’s a mechanic lying under a bus or a technician hanging from a utility pole. "The technology these people need is often much different from the types of technology that application developers created in the past. It is important to correctly reach out to these new types of users."
While IoT business models may be new, common lessons, challenges, and solutions can be applied across any emerging field in which best practices have not yet been established. Technology providers such as Cisco and Oracle are helping IT professionals understand what they need to do to make it all work. Manufacturing companies such as Emerson are paving the way with real-world applications. Dummann’s architecture team is creating playbooks to guide the business. The trick, he noted, is to develop guidelines without imposing rigid constraints on how each business unit moves towards its goals.
"The business is invariably more willing to adopt our recommendations if they are not forced upon them," he said. "EA has become the ringleader for developing these playbooks and creating these reference models, but we must work closely with the business to understand a whole new set of needs."