Bench Talk

The Internet of Things (IoT) offers a compelling vision of applications able to distill raw data from hundreds of thousands of sensors into meaningful information. For IoT developers, however, that vision becomes dulled by the reality of interconnecting the many local devices and cloud-based resources that make up an effective IoT application. Although standard mechanisms exist to create the foundation of an IoT system, the greater framework that ensures interoperability among the pieces of an IoT application remains weak at best.

For the enterprise, even critical issues such as performance, availability, and security take a back seat to interoperability. In the grand scheme of connected things, IoT developers who are struggling to integrate sensor devices with edge gateways and the cloud will have little time to create the kind of disruptive applications that enterprises imagine in their IoT strategies. For corporations expecting significant ROI for their IoT efforts, the cost of vertical integration can shift the balance sheet into the red. According to market-research firm Gartner, half of the cost of implementing an IoT application this year will be associated with integrating its component elements.

Worse, the challenges related to integrating the layers within an IoT application are only the beginning. Enterprise-level IoT applications will also require horizontal integration across multiple individual IoT application “silos.” For example, manufacturers looking for better efficiencies will need to interconnect industrial IoT applications with IoT applications from the supply chain, inventory, shipping, and more. In fact, market analysts at McKinsey say that 40 to 60 percent of the total potential value of an IoT application depends on the ability to achieve interoperability between different IoT systems.

Yet, ensuring interoperability even within a single IoT stack is not an easy task. An IoT system can be the very definition of a heterogeneous system, comprising a widely diverse set of devices and resources collected in its multiple layers. To work together, these pieces need a common framework of higher-level protocols and models for securely recognizing each other and exchanging information about capabilities, data, and status. While the Web’s standard HTTP protocol enables interoperability with a few request methods, header fields, and response codes, the IoT of course needs much more information not only about the nature of the interaction, but also about payload structure and semantics.

The diversity of devices and resources even in a single IoT application requires a more extensive set of standard methods for handling the possible interactions among the various pieces of an IoT system (Figure 1). Besides the World Wide Web Consortium (W3C), standards bodies and industry groups are each approaching the IoT interoperability problem from the perspective of their own charter and domain of expertise.

Figure 1: IoT systems comprise diverse devices and resources, requiring common APIs, models, and protocols to ensure interoperability. (Source: W3C)

Groups are even focusing on individual layers such as the edge devices, given their rapidly growing role in IoT systems. For example, efforts to define IoT edge-device standards are underway in the international ISO/IEC JTC 1/SC 41 Edge Computing group, the Edge Computing Consortium in China, the Open Edge Computing Initiative in the U.S., and even in specific application segments with the Automotive Edge Computing Consortium.

For developers, the essential question remains how well proposed standards address their requirements. In its own efforts within the W3C, the Mozilla Foundation is taking a more direct approach to test its proposals in the real world. Along with a proposed API for thing interactions (Figure 1), the group has recently provided a reference implementation of a compatible IoT edge device based on a Raspberry Pi. Mozilla’s Things Gateway project includes open-source software and a step-by-step procedure that developers can use to quickly implement the edge device. Mozilla’s approach helps fill a critical gap between standards and requirements in a rapidly evolving arena such as the IoT. Perhaps even more important, this approach provides a real solution for developers who need even a simple testbed to ensure interoperability of their own IoT device designs.