Remote Hardware Configurability for IoT Device Management: Will Vendors Finally Be Able to Provide Device Management That Delivers Energy Performance Improvements to IoT Hardware?

A core purpose of deploying a commercial IoT solution is to reduce truck rolls for businesses. Remote hardware configurability has the potential to directly improve truck roll–related performance metrics. It can form a core part of the value proposition for future IoT device management services. Remote hardware configurability is the ability to turn on or off certain elements of the hardware, such as low-power wide-area transmitters and receivers. The trend toward common standards in the IoT industry includes not only newer industry standards like the Lightweight Machine-to-Machine (LWM2M) protocol but also FIDO Device Onboard (FDO). The former makes it possible to have remote hardware configurability while the latter addresses concerns about security when deploying IoT devices that are expected to become increasingly sophisticated in their hardware configurability functions in the field.

For many device management customers, remote hardware configurability requires more than just hardware; it requires device managers to fulfill certain security criteria and for device managers to also offer group policies for connected things. The group policies facilitate the configuration of a group of sensors deployed in a certain area of the factory or a fleet of sensors (telematics units) attached to a group of automotive vehicles. Remote hardware configuration can be performed on multiple devices instead of manually on each individual device. To deliver these services, a sophisticated over-the-air update mechanism is important since (1) it allows security fixes to be provided and (2) managing a mid–device firmware upgrade failure often requires reverting the firmware and system files back to the previous version. These are all core functions of a device manager, and therefore device managers are vital for facilitating remote hardware configurability in bulk and in maintaining a secure environment.

Chipset vendors such as Qualcomm have started offering chipsets that are LWM2M compliant, and this not only simplifies both interoperability and the onboarding process but also introduces the potential for remote hardware configuration on an entire fleet of devices. LWM2M is a key driver in accelerating hardware configurability in certain industry verticals like utilities and energy (smart meters and water leak sensors) that can run on a datagram transport layer security network protocol. Also, since LWM2M-compliant chipsets have device agents natively built in (See “IoT Solutions Transition from Proof-of-Concept to Deployment: How Device Agents Enable Critical Device Management Services,” IN-6301), it is possible to leverage those device agents and configure hardware level settings remotely by switching the on or off components within the sensor (such as transmitters or receivers) to preserve battery life.

Implementing remote hardware configurability in IoT deployments requires more than just hardware. The benefits can only be leveraged once a device management platform is ready to deliver remote configuration from instructions on the user’s dashboard to the device itself. Most importantly, remote hardware configurability gives customers the flexibility to optimize their IoT deployment by balancing energy efficiency with performance requirements. This gives the customer greater control over their IoT fleet and provides them with flexibility to adjust the fleet’s hardware regarding data-driven insights, which are generated through predictive analytics.

Remote hardware configuration has been held back in resource constrained environments due to energy constraints caused by cloud computing since a greater amount of data transfer to a server for running analytics would lead to greater use of network connectivity and accelerate energy consumption on battery powered devices. This hurdle has been overcome in large part due to edge-computing environments with an analytics framework such as Tiny Machine Learning (TinyML) that can function on low-compute devices. Since advancements in edge compute now go beyond business rules (i.e., it’s more than a contingency rule when device connectivity with the cloud falters) and includes performing analytics with a lower latency response at the edge instead of higher latency cloud-based analytics applications, it becomes possible to leverage remote hardware configuration. This is because the analytics are performed at a relatively lower cost to energy consumption than before, and so TinyML addresses the energy constraints and the analytics have to be implemented, making remote hardware configurability a useful feature in avoiding truck rolls.

To conclude, remote hardware configurability has historically faced hesitancy from hardware manufacturers (including module and gateway original equipment manufacturers). This hesitancy has been confirmed by numerous research interviews conducted by ABI Research’s IoT Network & Services Research Service. The main motivation behind hardware vendors’ preventing remote configuration of hardware is an overriding concern of security—especially fears that a hostile actor could exploit a remote hardware configuration mechanism and execute malicious code. These concerns are being addressed through access control that helps to prevent hackers from leveraging one compromised device to access other devices on the network. While FDO addresses concerns about knowledge-based secrets (passwords) being compromised and about device onboarding, the role of secure boot and trusted platform modules requiring a hardware root of trust while remotely configured devices have an altered boot sequence remains an open question in order for hardware configurability to truly surmount hurdles.