The Internet of Things (IoT), with its ability to create vast amounts of previously impossible-to-collect sensor-based data from physical objects, creates both an opportunity and a challenge for enterprises. The opportunity arises from the chance “to improve the efficiency of their operational business processes.” The challenge arises from the fact that traditional business process management (BPM) systems were not designed to deal with the huge quantity of distributed data that needs to be extracted and analyzed automatically in real time.
The authors propose joining distributed analytics (DA) with IoT and BPM to meet that challenge. Working together, the three distinct technologies can combine “to drive business processes automatically with information [derived] from the physical world.” IoT “couples the physical world to the digital world” by allowing individual physical objects to not only generate data, but also be digitally actuated (when appropriate). DA delivers the necessary software intelligence to make business sense of the IoT data. This translates the raw data into information that BPM uses to drive and enrich business processes, and those processes can then deliver more value to the enterprise.
Converting IoT data into useful information can be a very difficult and complex task for many reasons, for example, the data itself is usually “heterogeneous in format and large in volume,” and the IoT resources can be “large in number [and] heterogeneous in nature,” as well as being owned and managed by diverse business parties. To solve these problems, the authors propose what they consider to be “a novel combination of existing [DA and BPM] technologies.”
The heart of the strategy is the use of containerized microservices. For example, the data concerns are resolved by bringing the analytics in a distributed format to the data rather than bringing the data to the analytics. This approach uses “parameterized microservices that are packed into software containers,” which enable them to be “dynamically deployable from a service repository into the IoT edge.” The number of IoT resources issue is addressed through the automation of “the deployment and management processes of the containerized microservices using a BPM engine.” The authors describe a proof-of-concept prototype implementation to demonstrate the feasibility of their approach.
For any reader who is exploring the possible use of BPM to deal with IoT-generated data, this paper is a must-read. It is a seminal work that should influence the subject for years to come.