After big data and machine learning, IBM is now turning to the Internet of things to rejuvenate its business. It's a good fit for IBM's revamped view of itself, but IoT is bigger than what IBM's current plan encompasses, and any long-term plan to embrace IoT would be wise to acknowledge that.
IBM's plans consist of a newly formed business unit dedicated to IoT, with a subsection of its Bluemix PaaS devoted to IoT processing and a whole new cloud service, the IoT Cloud Open Platform, to back it up. All of this fits logically into IBM's revamped business model, where services and APIs form the meat of their offerings.
This IoT push also links up with other parts of IBM's new business model. For one, the new IoT business unit also appears to be aiming for customers from many of the same verticals targeted by IBM's alliance with Apple. As intriguing as the pairing has been, it's so far yielded closed-ended offerings, with little appeal to the broad swath of midsized organizations that might generate the kind of large-scale demand IBM needs.
Also, a key part of IoT in its most forward-thinking incarnations involves the different varieties of data aggregated from different sources. Some of IBM's recently announced data services seem an appropriate fit for this kind of processing, though not yet in their current form.
Watson Analytics could work in this vein, with a data-cleaning function that could come in handy for processing data from heterogenous sources. But right now that service only works with static data sets, not the stream-based processing most commonly associated with crunching IoT data. That said, IBM's been assembling a galaxy of services (including this one) that mesh well with IoT-style processing. Bluemix offers a time-series database based on Informix, and IBM's work with Twitter also shows how real-time stream processing from arbitrary sources could be further integrated into Watson.
IBM's overall intentions aren't hard to parse. The company faces tough times as it sheds business units it's been traditionally associated with, and it's been scrambling to redefine itself as focused on big data, cloud computing, machine learning, and a bevy of other related ventures, all delivered as services and APIs.
IoT fits well into all of that, since much of what IoT is about is in IBM's current technological wheelhouse -- collecting lots of data from many sources, processing it in novel ways, and repurposing that data. But IBM is not yet speaking to anything beyond that basic, nonautonomous definition of IoT. IBM's announcements around this IoT initiative revolve mostly around analysis of data harvested from sensors and devices, rather than the machine-to-machine, smart-systems, or ad-hoc device interactions that characterize where IoT is headed.
Still, IBM's centralized service approach will likely appeal to enterprises where centralized processing is the order of the day. What's worth watching out for is whether the IoT Cloud Open Platform -- whatever that shapes up to be -- also comes to include IoT's more autonomous sides.