There's no easy way to do IoT management

The complexity of mobile's early days are a small taste of what IT will face in managing the internet of things

There's no easy way to do IoT management
Credit: Thinkstock

The pitches have already started: Various vendors promise to help you manage the tsunami of internet of things (IoT) devices coming. But beware: IoT management will be very difficult for a long time, and it may never reach the steady state of mobile management.

Potential good news: You likely won't have a flood of IoT devices coming from all quarters as we saw in the early days of mobile BYOD. But what IoT devices you have will be very hard to manage.

Remember when there were iPhones, Android smartphones, BlackBerrys, WebOS phones, Windows Phones, and legacy Windows Mobile devices? IT shops freaked out at all the variations they'd have to support.

Imagine an IoT world where thermostats, door locks, light switches, elevator stop regulators, smart glasses, Bluetooth internal-location sensors, comm badges, security cameras, heat sensors, conference gear, alarms, and the gazillion other devices and sensors that might exist in the products you use and environments you manage -- as well as the possible IoT devices and sensors in health care, transportation, oil exploration, logistics, retail, and public safety. In comparison, the half-dozen mobile devices that so freaked out IT look as frightening as a sleeping cat.

The mobile industry coalesced to two operating systems: iOS and Android. Both use APIs that have high overlap, so management vendors can now let IT manage all these devices from a single panel of glass using a consistent set of controls. The variability and exceptions are thus now quite manageable -- even if you add Windows 10 and MacOS computers to the equation (they use similar APIs).

We won't see that level of consolidation in the IoT world. Even if every type of device ended up being dominated by one or two providers, the huge diversity of devices would still mean hundreds of providers. The chances of them agreeing to a common set of APIs is close to zero.

It won't be total chaos, of course. Some industry segments will get picked up by systems management vendors as critical mass is reached. Essentially, vendors within a product class and/or industry must standardize that segment to the point where management providers like VMware, MobileIron, Soti, BlackBerry, IBM, and Microsoft can all support the major IoT segments' offerings.

But that's not easy because standardizing means generalizing key aspects of the product, reducing the differentiation that vendors want to keep customers locked into their products. That's usually a last-resort decision when either of the following occur:

A dominant provider essentially forces compatibility

Apple did that in its 2010 iOS APIs, leveraging its dominant position in the enterprise for non-BlackBerry devices. Google quickly followed suit for Android, whose enterprise adoption has been tiny -- Google basically hoped for Android devices to come along for the ride by adopting iOS-like APIs. Apple is now using the iOS APIs so that managing IoT devices like the Apple Watch and Apple TV becomes a simple extension to managing iPhones and iPads.

One example in IoT includes common management for Android-based smart glasses from several manufacturers (Atheer, APX, Intel, ODG, and Vuzix), which VMware AirWatch this month will support from its management tools. This could become one segment standard.

A fragmented industry won't move forward without standardization

The best example here is Wi-Fi. Until the Wi-Fi Alliance, a vendor group, came up with the Wi-Fi wireless compatibility standard, businesses refused to adopt wireless LANs in any volume because pre-Wi-Fi devices could only talk to other devices from the same company, though they all used the same 802.11 transport protocol.

One example in IoT is the Data Distribution Service, a set of APIs and data transfer protocol from the Object Management Group. This lets various manufacturers' devices talk to various manufacturers' analytics back ends. 

We'll see both patterns occur in IoT.  But it will take years for the shift to either approach in each domain. In the meantime IT will be faced with a grim reality: Most IoT products will have their own security models and management tools, creating an unwieldy mess for IT. That'll slow IoT adoption out of necessity.

In the industry standardization game, someone has to blink first, but the stare-down will go on for years until someone finally has to back off. In the meantime, pick and choose your IoT initiatives carefully -- you'll be able to handle only so many of them.

From CIO: 8 Free Online Courses to Grow Your Tech Skills
View Comments
Join the discussion
Be the first to comment on this article. Our Commenting Policies