Making IoT Applications Enterprise-Friendly

Jan Höller and Senthamiz Selvi A
September 16, 2019


The reach of IoT applications is into nearly all sectors of the global economy. The total transformative economic potential of IoT has been valued at up to 11 TUSD per year by 2025 [1].  However, the value is spread across a spectrum of applications from different sectors.

As can be understood, this diversity creates challenges in how to implement and employ the range of IoT applications, both in costs and in time to market but moreover how to effectively integrate them into the business environments of any and all enterprises. The current predominating practice of building IoT applications is to start with a single problem related to the “thing” or a machine and build a bespoke solution for it. However, the traditional bottom-up Operational Technology (OT) problem approach to an IoT solution benefits from being complemented with a top-down enterprise business process approach.

Maturing Towards the Top of the IoT Technology Stack

Turning to the IoT technology stack, one can in a sense say that the technology evolution via research and innovation has over the years gradually been sliding its focus higher and higher up the stack. From the use of IP, networking, and connectivity, to the cloud, further to the edge, and in recent years to AI. This progression is natural as firstly, the raison d’être of IoT is about insights and automation in the change of process and operational models of enterprises, and also about new business models. Secondly, and simply, technology’s usefulness is gradually maturing higher and higher up the stack.

The IoT operational and delivery model is via IoT application enabling platforms that provide horizontal tools for building applications. To date, those platforms have primarily been about infrastructure services, like connectivity, protocol adaption, device management, and secure device onboarding. The more data-centric aspects of the platform are to a large extent still rather simple and point-application focused, like visualization of insights, simpler automation tasks based on rules, or applying a specific AI-based task like predictive analytics to understand when and how a machine might fail. But what is remaining is still how such point applications can be effectively integrated and enriched into the more complex overall enterprise business processes.

As an example, the aforementioned machine failure prediction is just one event that is part of predictive maintenance, e.g. a machine cutting tool wearing out might actually be detected post-production as a different event. In the bigger picture, servicing a production machine has an impact on production plans including customer risks that need be folded into the equation, so does the availability of spare parts and service engineers, and the production line operation might need to be adjusted to shift the time for machine service to a more optimal time, all requiring the proper planning and adaptive processes, see figure 1. The machine-centric bottom-up approach to the IoT application clearly needs to be complemented with an enterprise-centric top-down approach, and the solution will also rely on a larger set of diverse AI tools, [2], all that need to interact with the enterprise systems like Enterprise Resource Planning (ERP) or Customer Relations Management (CRM).

Figure 1: A predictive maintenance scenario.

Figure 1: A predictive maintenance scenario.


A Complementing Enterprise-Centric Process Approach

Our approach to adding an enterprise-centric perspective to IoT applications can be summarized by the following main points.

  • Identifying high-level enterprise requirements, and modeling a top-level business process flow in an end-to-end enterprise-centric application context.
  • Further detailing of sub-processes to the level of reaching individual IoT resources and tasks that can be mapped to microservices, thus being application-independent and ideally reusable, [3]. This is a recursive activity at different interconnected levels of abstraction thus bridging the design time and run-time phases.
  • Mapping of the most granular sub-processes to functional components that can be represented by underlying IoT resources or microservices.
  • Orchestration and execution of both enterprise process relevant events and IoT event and data streams in a consistent framework of reusable components.

Our work is exploring the use of Business Process Model and Notation [4] for the enterprise process modeling part. BPMN was originally not developed to include IoT applications that involve industrial processes or OT equipment, but the exploration of applying BPMN to IoT is not new. BPMN is a promising baseline that can provide a missing link in making IoT applications enterprise-friendly but would require further development. We provide a summary of related work and point to new areas where we believe exploration and further work is needed, and the target is primarily evolving BPMN itself.

To date, the following challenges have been dealt with that either specifically address IoT, or are general and applicable to IoT.

  • Proper exposure of IoT resources as process resources in standard BPMN models, see e.g. [5] and [6] for a discussion. The introduction of IoT resources that capture real-world events and actions introduce a degree of unreliability that must be properly taken care of when modeling processes that involve IoT devices. It is important that the process design is done with clear objectives and boundaries to avoid making the process too complex and cumbersome. Extensions to BPMN to include IoT resources as process resources have been proposed, [8].
  • Stream processing for BPMN: IoT provides enormous amounts of real-time data in the form of a stream of events. The benefits of integrating these real-world data to existing business processes are valuable in making real-time decisions and optimizations. BPMN lacks abstraction mechanisms to encapsulate event streaming in the current specifications. Event Stream Processing Units (SPUs) has been proposed as an integration concept for event stream processing to integrate with business processes, [6].
  • BPMN execution and microservices orchestration: Achieving scalability becomes possible when the right level of abstraction coupled with dynamic orchestration is involved in process modeling. When combined with the right engine, BPMN makes it easy to connect tasks in a workflow to microservices and to do so in a way that doesn’t violate the principles of loose coupling and service independence. [7]
  • Extension of the BPMN Lane to include IoT resources [8]: The introduction of a new Sensing Task with native software components referencing the IoT Domain Model of the IoT-Architecture as a combination of swim lane and process activity-centric resource model. Integration of the semantic model as parameters to the new task, extension and practical testing of the graphical model of a business process modeling tool is presented.

Given the above progress and considering further anticipated needs for applying BPMN for IoT, we also propose the following new areas to explore.

  • Separation of business process event and logic from IoT data streams. This implies an association but separation of enterprise event to underlying microservices that produce and consume IoT event streams, e.g. model training on IoT data, inferencing, etc. that will become continuous background activities separated from the enterprise process events. A BPMN engine can then be used for both the orchestration of microservices and the execution of the enterprise events separate from the underlying IoT event streams.
  • As IoT processing generally is distributed in physical space, enterprise sub-processes will benefit from being distributed too for different reasons. This consideration needs to be taken into account throughout the modeling process from the top level of abstraction to the distribution of microservices across e.g. a cloud and factory shop floor. This will require extensions in modeling to capture requirements and diverse constraints, both from an OT perspective and from a distributed cloud and edge computing perspective.

BPMN also has the capability to model the interaction between a number of different participants and hence lends itself well to explore IoT-centric processes that span across value networks of collaborating actors. The separation of business events from IoT event streams would also allow an interconnection across actors at those different levels as needed.

Figure 2: A conceptual framework to bridge enterprise processes with IoT.

Figure 2: A conceptual framework to bridge enterprise processes with IoT.


Our approach to harness the power of IoT enabled business process modeling is conceptually summarized in figure 2. If we take the example of Predictive Maintenance use case by taking into consideration a top-level business process (Level 0) as in Figure 1, this can be further detailed to subsequent levels depending on the decomposition of subprocesses involved in the higher level flows, and eventually orchestrate the underlying  well defined microservices in the IT infrastructure layer. Level 0 of the business process will depict the flow of events at 30,000 ft. view of the system integration involving both the IT and OT functions of the manufacturing organization. Level 1 will further decompose the process into subprocesses of specific aspects of the organizational Lanes and so on. As one drills to the atomic activity (say Level 5) which has the implementation service hook to the underlying  IoT resources and microservices, we observe the full potential of seamlessly bridging the enterprise perspective with the IoT perspective covering both the design, orchestration, and execution phases.


In order to capture the full potential of IoT across different applied sectors, means for effectively developing and integrating IoT applications that consider the richer and wider enterprise needs are still missing, but baseline work exists that can be further exploited. Our approach builds previous work, and we propose further work that can help close the gap between the enterprise process-centric and IoT-centric perspectives that we believe will make IoT applications more enterprise-friendly.


  1.  “The internet of things: mapping the value beyond the hype”, McKinsey Global Institiute, 2015, available from
  2. J. Höller, V. Tsiatsis, and C. Mulligan, “Toward a Machine Intelligence Layer for Diverse Industrial IoT Use Cases,” IEEE Intelligent Systems, vol. 32, no. 4, pp. 64–71, 2017.
  3. “Accelerating Industrial IoT Application Development through Reusable AI Components”. Senthamiz Selvi Arumugam, Ramamurthy Badrinath, Aitor Hernandez Herranz, Jan Höller, Carlos R. B. Azevedo, Bin Xiao, Valentin Tudor, “Accelerating Industrial IoT Application Deployment through Reusable AI Components”, in proceedings from IEEE GIoTS 2019.
  4. Business Process Model and Notation, available from
  5. Haller, Stephan and Carsten Magerkurth. “The Real-time Enterprise: IoT-enabled Business Processes.” (2011).
  6. Appel, S., Kleber, P., Frischbier, S., Freudenreich, T. & Buchmann, A. (2014). Modeling and execution of event stream processing in business processes. Information Systems, 46, 140–156.
  7. “BPMN and Microservice Orchestration”
  8. Meyer, S., Ruppen, A. & Magerkurth, C. (2013). Internet of things-aware process modeling: integrating iot devices as business process resources. In International conference on advanced information systems engineering (pp. 84–98).


Jan HollerJan Höller is a Research Fellow at Ericsson Research where he is responsible for defining and driving the IoT technology and research strategies, and also to contribute to the corresponding corporate strategies. He established Ericsson’s research activities on IoT over a decade ago. Jan is a co-author of the book "Internet of Things: Technologies and Applications for a New Age of Intelligence" that was recently released in its 2nd edition. He has held various positions in Strategic Product Management, Technology Management and has, since he joined Ericsson Research in 1999, led different research activities and research groups. He has for a number of years served on the Board of Directors at the IPSO Alliance, the first IoT alliance formed back in 2008. Jan currently serves on the Board of Directors of the Open Mobile Alliance and is a chairing the Networking Task Group in the Industrial Internet Consortium. He is a frequent speaker at industrial and academic conferences and events.


Senthamiz Selvi ASenthamiz Selvi A is an Experienced Researcher at Ericsson Research, India, currently focusing on the areas of the Internet of Things and Artificial Intelligence. She is an HFI-Certified Usability Analyst. Selvi joined Ericsson in 2009 working on the Business Support System portfolio products as a Software Architect before moving to the corporate research unit. She received her bachelor's degree in Computer Science and Technology from Usha Mittal Institute of Technology, Mumbai. Her current research interests include IoT, Machine Learning, Distributed Ledger Technologies and Business Process Modelling.