IoT device management is becoming as hot a topic these days for enterprises as mobile device management became over a decade ago. Why? According to Dave McCarthy at Bsquare, enterprise IoT “proof of concepts” are maturing in to larger and more distributed deployments given the success of early tests.
“We’re now seeing massive adoption in the marketplace, with not just hundreds but thousands, tens of thousands and even hundreds of thousands of sensors being connected to manage everything from smart buildings to smart factories,” McCarthy said.
IoT device management introduces new challenges to IT and OT teams, related to scale, security, connectivity, and compute. While they have significant experience in locking down mobile devices and network infrastructure, being able to now also be responsible for connected systems – lighting, security, cameras, HVAC and more – is driving a new focus on how IoT and IIoT solutions can be integrated into their existing management platforms or connected on all new platforms.
“It’s important to look at the challenges holistically,” McCarthy said. “These initiatives have evolved into much more than just smart, connected equipment in the field. The implementation of IoT means that the data generated by sensors must also be taken into consideration along with the physical devices themselves.”
Some businesses have tried to build their own device management solutions for IoT, but have encountered problems, including getting budgets approved which include major fixed costs. McCarthy says the programs succeeding are opex- vs. capex-based, with cloud-delivered solutions.
“Bsquare is collaborating on IoT device management with AWS,” McCarthy said, and identified four reasons why enterprise IoT projects are not getting off the ground:
“Another challenge we’re solving is the onboarding of large numbers of devices in disparate locations,” McCarthy said. “Automation of this process, in addition to systems which continually and securely measure device health and performance, is critical. By using templates and workflows, devices can be onboarded, configured, and put into the right groups without human intervention.”
“All major cloud providers who are working hard to support the IoT industry are realizing that, while the cloud is great as an aggregation point for data and scale, there are limitations for resource constrained devices or devices needing to react to real-time conditions,” McCarthy said. “This is why they and other service providers are investing so much in edge technologies, including compute closer to the edge or at the edge.”
McCarthy is seeing a continued and fast evolution between centralized and decentralized computing, and the need for a mix of networks and standards. “There will never be one standard – we will always have a mixed environment – a combination of things – which is why it is so important to adopt an enterprise strategy that includes management software on devices. True device management cannot happen without this, nor can the optimization and securitization of the data sent by signals from those devices.”
McCarthy also sees more complex device management surfacing, “where software updates are important, security updates are important, and over-the-air must be supported. Devices are more complicated now. There are touch screens and other features so we must move device management way beyond configuration into supporting these unique needs.”
McCarthy called the world of IoT device management and data management “married” – particularly given the trend of sharing data with different systems.
“We’re seeing a mix of equipment manufacturers, owners or renters of that equipment, third party repair and service agencies, all whom benefit and can bring value when devices send them alerts, notifications and other information. One source of data with many different stakeholders? It’s a big challenge for the industry as there is no way to manage all this in a unified manner today,” McCarthy said.
“The OEM is driving a lot of this, with new connected products that don’t just collect data, but share it – whether they charge for it or not, they are creating value,” McCarthy said. “Depending on how much value can be generated, there is definitely a new connected device economy being created. Without a doubt, if you believe in any of the numbers around scalability required for IoT in general – you cannot use the old method of staring at dashboards and lights turning red – automation as part of device and data management is imperative.”
McCarthy talked about the level of automation and its role in real-time or near-real-time solutions.
“Depending on the level of confidence enterprises need to detect, remediate and report – and depending on how much they want closed loop systems to, for example, turn off a machine when it reaches a certain temperature or other policy – consideration needs to be given to the economics associated with best efforts on one end, and ultra-low latency on the other. For the latter, edge computing is going to be required, with local networks sending data up to the cloud for any centralized administration applications.”
When there are real-time process controls, McCarthy said this can impact safety issues. Latency caused by a cloud approach won’t work. “We work with subject matter experts on advanced projects like these—those who understand the machines, the relationship of humans to those machines, the value of keeping those machines running and keeping those people safe—to get to a true ROI for investing in augmented systems where lives are at risk.”
It may all come down to data science, McCarthy explained. “Data science approaches are invaluable when planning an IoT project that includes device and data management requirements. Getting the subject matter experts in the business to collaborate with IT and OT in the planning, implementation and ongoing management phases is golden.”