When designing products and systems for the Internet of Things (IoT), manufacturers often tend to focus on only a single aspect of the process: embedding wireless communications, or establishing a cloud connection, or writing web or mobile-based software to control an IoT device.
Taking an isolated approach to the IoT, however, is a huge mistake. Enabling security, performance, reliability, in-field updates and long-term maintenance requires a cohesive, end-to-end approach—both as a design paradigm and when planning to extract the real value of the IoT, through analysis of the data generated by connected devices.
The Importance of ‘Things’ in the IoT
Most often, the focus in IoT is on the “Internet” part of the phrase. But it’s the approach to the “Things” that is crucial for designing an IoT product or system, industrial or otherwise. Effective implementations require expertise and experience in the full IoT spectrum—from the communications chips embedded in the connected device, through cloud computing and network security, to control via a web-based or mobile application.
"Most often, the focus in IoT is on the ‘Internet’ part of the phrase. But it’s the approach to the ‘Things’ that is crucial for designing an IoT productMost often, the focus in IoT is on the ‘Internet’ part of the phrase. But it’s the approach to the ‘Things’ that is crucial for designing an IoT product"
Plus, every piece of IoT technology needs to work seamlessly with all the other pieces to create an integrated, secure, end-to-end system. Parsing responsibility for different pieces to different entities, and assuming that the correct integrations and handoffs will occur, is a recipe for disaster.
It’s also important to have a clear idea of the ultimate purpose of any IoT implementation, and to design all aspects of the product or environment so they work together to achieve that purpose. For example, what is the mix of legacy and new devices that need to be taken into consideration? What kinds of users will be interacting with the system, and how should the devices and applications be controlled? What level of responsiveness and availability is required?
Design Considerations across the Full IoT Spectrum
IoT solutions must encompass connected device, cloud and control application technologies. Here are a few of the capabilities and kinds of technical expertise required for each of these areas:
For connected devices:
• Internet connection providing reliable, secure cloud connectivity
• A tested, provisioned, up-and-running IP stack
• A scheduler that is debugged, tested and production-ready
•Wireless connectivity (in industrial settings where wireless connectivity is both possible physically and advisable) that can handle addressing, connectivity to the network and to other devices, and security
• Device addressability, identification and authentication
• Methods for distributing and managing software updates, upgrades and bug fixes remotely and over time
For cloud connectivity:
• Tight network security, including access control; device and cloud authentication; data privacy; data security both at rest and in motion; firewalls; wireless encryption; and continual updating to the latest security standards
• Networking protocol support; in the case of IoT environments with legacy equipment, this means supporting both old, technically obsolete protocols as well as future protocols that have not yet been designed
• Reliability and resilience, including for scheduling, automatic responses, and what happens in an IoT environment if the cloud connection is lost
• Scalability; it’s one thing to offer a given level of security, performance, reliability or responsiveness for a handful of connected devices, but what happens when it’s time to scale to thousand or millions of devices?
• Cloud service architecture optimized for chosen cloud service providers, each of which offers different sets of services
• A data infrastructure able to handle and process the constant stream of data generated by always-on devices
For application control:
• The type of application control: web-based, mobile app or other
• Secure connectivity of the control application with the cloud
• Identifying and authenticating authorized users
• Discovery, selection, authentication and management of wireless LAN access points
• Management of user interactions with schedules, sensor input, actuators, data generation and analysis, and other automated activities
Taking a Platform Approach
It is highly unlikely that the owners of an IoT implementation also are experts in all aspects of IoT connectivity— especially given the recent emergence of IoT technologies and the long histories of most industrial and commercial equipment being connected to the IoT. Expecting in-house teams to build their own IoT solutions is unreasonable and unwise. In most cases, it is just too difficult, expensive, time-consuming and risky.
A smarter approach is to start with an IoT platform that already integrates all the technologies—across devices, cloud and application control— required for a fully functional, secure, reliable IoT solution.
An effective IoT platform should offer all of the functionality needed to deliver this top-notch level of IoT connectivity, including end-to-end standards-based security. Through data analytics, it should make it easy to track IoT products’ real-work performance; improve everything from time to market to maintenance costs; and even deliver insights into how to improve the design and functionality of future IoT products and systems.
And because the very concept of IoT connectivity is at odds with the existence of silos, an IoT platform should be able to operate as part of a robust ecosystem— integrating with third-party clouds and keeping pace with evolving standards, protocols and security updates.
No one can see into the future to know what features, technologies or intangibles will be important for the next generations of IoT solutions. For that reason, any IoT platform chosen today must adapt easily to changing innovations, standards, components and demands. Flexibility enables future-proofing—which allows IoT environments to remain viable without the expense and hassle of ripping and replacing each time a technology advances at the device, cloud or control application level.