Identifying the Major Technological Domains in an IoT Architecture
To deliver an end-to-end IoT solution, architectures will potentially require seamless interoperability across the six following technology domains. These domains are illustrated in Figure 1:
Figure 1: High-level domain design for an IoT architecture [Source: Machina Research, 2015]
- Device domain – connected assets including sensors, devices, and modules
- Local Network domain – connectivity technologies enabling internal transfer of data from sensors and devices to other devices or a local network gateway
- Wide Area Network domain – connectivity technologies enabling the transfer of data directly from devices or local network gateways to external Service Enablement domain
- Service Enablement domain – platforms and middleware
- Applications and Data domain – provisioning, development, storage, and management of applications and data
- Enterprise Systems domain – back-end enterprise / corporate systems
Each domain is comprised of a specific set of products, services and skills. Within IoT, the configuration of these domains may change from use case to use case. Given this characteristic, one of the crucial considerations for enterprises is to identify the tools and enablers which make implementing IoT solutions across these domains as easy and simple as possible.
Removing the Interoperability Barriers
Standardisation is one important way forward to removing the interoperability barriers between (and within) the various domains. In standardisation, technical standards are defined and enable agreed interoperability and compatibility workings for data and communications between and across devices, connectivity technologies and platforms. The ‘language’ has been agreed.
Another approach which reflects the early stages of M2M and IoT is that of enablement platforms such as ThingWorx designed as being device and connectivity agnostic. In this approach, platforms are designed with a wide range of interoperability tools and approaches to allow for the smooth interoperability between devices, connectivity technologies and platforms. This approach is fundamentally based on pre-configured technical integrations between assets. As the diversity of assets remains fairly ‘manageable’ and ‘predictable,’ assisted by being implemented within defined sectors or segments, platforms that are device and connectivity agnostic deliver significant shorter to medium benefits. In the longer term however, standardisation has to be the way forward for massive heterogeneous asset implementations.
Modular Approaches will be Important for IoT to Grow
Compared to M2M where the full stack across all domains was designed and implemented as one, IoT will benefit from a highly modular approach to its technology architectures. With standardisation being difficult and still some time away, enterprises should recognize that to ability to address the ‘anarchy’ of devices and connectivity technologies, as well as dynamic shifts over time, data types, and applications will be crucial and that no single architecture will fits all requirements. IoT will be best designed according to the principles of Lego building blocks, modular, quick to construct, rebuild and add on to.
This extends not only to assets working with one another but more importantly, to the languages and programming skills that will be required to develop and enhance the applications and data structures behind each IoT solution.
For more insights around Blueprints and Feasibility Studies for Enterprise IoT, read the Second Machina Research White Paper.