So, you've decided to pull the trigger on an Internet of things (IoT) project. You've read the stories about how it can drive new efficiencies into your business and perhaps even open up new revenue streams. Be prepared for the IoT world's dirty little secret, though: managing these devices is like herding cats.
Everything starts off well enough on the average IoT project. The technology team tinkers with a few devices for a proof of concept, installing and testing the necessary software and producing something that impresses senior management. Things might even work out well enough in a pilot project with a dozen or so devices running on non-critical equipment in the field.
Get ready to scale
Sooner or later, though, IoT projects have to scale to generate their promised benefits. You have to install enough smart lighting across your business to generate worthwhile energy savings. You must equip enough trucks with connected tachometers and GPS systems to properly monitor your fleet, or you must install enough water flow sensors to generate meaningful information about your industrial plumbing system. Whatever the application, IoT devices quickly multiply until there are thousands in your infrastructure. The management techniques you use for your IT network just won't do.
You will need an IoT management system to help you control this vast array of devices, not least because they aren't as powerful as enterprise IT equipment. Many of them are constrained, with small energy footprints, running bare-bones operating systems. They will often be in places that an in-house technician won't be able to visit, making remote management especially important.
IoT management systems often run in the cloud, from which they connect directly to IoT gateways that in turn manage devices in the field. The hyper scale cloud service providers all offer them. Amazon has AWS IoT and Google offers its own competing platform. There's another option on Microsoft's Azure cloud, and IBM manages IoT via the Watson IoT platform. There are a variety of other management platform providers, too, such as Bosch, which offers its IoT Suite.
There are several facets to IoT management, and you should be sure that your chosen solution covers them all. The most visible task for many project teams, especially in the early stages, is device management. This extends throughout the entire device life cycle from provisioning through to end-of-life.
Devices should be provisioned using unique digital certificates, enabling them to prove their identity on the network and avoid rogue units that could compromise the infrastructure. This will also enable them to encrypt their communications. Those certificates should be revocable at any point, protecting the organisation from device theft and making decommissioning easier.
The IoT management system should enable administrators to monitor devices en masse, quickly sorting and filtering them, and should also handle group configuration and mass rolling firmware updates to keep all devices current and secure.
Beyond the devices themselves, consider connectivity management. These systems may connect to the cloud in several ways, including 3G or 4G cellular networks, or other low-powered wireless solutions such as Wi-SUN or Lora. Something must manage those connections, which may span multiple carriers depending on your regional scope and commercial needs. That means managing the SIM connectivity and contracts, along with bandwidth usage and possibly fallback connectivity.
With your IoT infrastructure sorted out, you still have to manage the data that it generates. IoT devices churn out a lot of information that will present new technical challenges. Storing and analysing it will require new storage mechanisms and database solutions. NoSQL data structures are often more appropriate than traditional relational systems here. In some cases,you might need to process that data in real time, creating an additional need for streaming analytics expertise.
This is all doable, but it's important not to dive in without mapping out these challenges and solutions first. By all means start small. That's crucial because you are almost certain to learn from early mistakes. But you should also think big, creating the capacity to smoothly manage a rapidly growing IoT infrastructure. When you're faced with the task of managing a constellation of tiny assets spread across a large area, it's vital to think ahead.