Dell put a stake in the Internet-of-Things (IoT) ground last week at a large analyst and press event in New York City. Hoisting its own IoT branding, the company introduced a bespoke moniker — IQT — to describe its efforts. IQT does not explain itself at first viewing. It requires a tag line or some other clarification. It stands for, essentially, Intelligent Quotient of Things, which is still pretty opaque, but the meaning is “intelligent Internet of Things.” Never mind that the Central Intelligence Agency is already using IQT for its own purposes …
The whole team was there, from Michael Dell on down, and the message was pretty consistent. If you didn’t get it after two days, something was definitely wrong. Given the still-fluid-but-rapidly-solidifying boundaries of the IoT that I described in a recent article, it’s clear that Dell will not be addressing the market for the trillion sensors that will represent the endpoints of IoT networks. That business will go to one or more of the firms that work with ARM architecture, which is more suited to low-power applications. A major component of Dell’s strategy is the IoT gateway (server) that Intel is backing.
And the question arises as to whether that gateway business will be big enough to support a major portion of Dell’s future market position. I sat down with Joyce Mullen, the senior vice president and general manager of the company’s Global OEM and IoT Solutions Group, and she gave me a sense of how it’s supposed to work.
From Dell’s perspective, the company can address three areas of IoT topology: the aforementioned gateway, something called the “distributed core” (more on that in a minute), and the cloud. The cloud is no problem. Dell is already a major cloud supplier from a number of standpoints. It sells lots of OEM hardware to the large public cloud suppliers. Through its Infrastructure Solutions Group (made up mostly of business units from the former EMC), it sells piles of hardware, software, and services to enterprises for cloud installations. And its “converged infrastructure” products (configurable units with storage, networking, and compute elements) carry a growing part of this load.
To get a sense of the distributed core, imagine a manufacturing firm in the dairy products business (one of the areas Dell’s IoT efforts have already had some success, albeit at the retail end). At the lowest level, on the machines and pallets in each factory could be tiny, low-power temperature sensors. These send out a constant stream of numbers, most of which aren’t interesting. On an increasingly crowded network, there is no way that this information should flow very far. So, a gateway — say one for each line — could receive this information and decide how interesting the numbers are. If the temperature rises past some level, the gateway might throw an alert, perhaps to the distributed core element, which is a fully empowered server at the factory level that can do highly intelligent things like turn up the freezer, slow down the line, or call in service for a machine.
The distributed core element could also see whether other lines are experiencing the same problem and begin to isolate the fault. And in the best of cases, it could use machine learning and artificial intelligence to spot the rising temperature pattern before it hits a critical limit (that’s the IQ part). Finally, it could also refer back to the cloud — which might be a server farm at headquarters or hosted by a cloud provider like Amazon — to compare notes with other distributed core elements in the company’s other factories.
Now, Mullen pointed out that where the actual intelligence lies, where the decision-making is done, and to where the data will flow are all up for grabs. Some installations may do most of the work in the gateways, some, in the distributed core, and still others, in the cloud. And not all installations will have all elements. Clearly, in applications where speed and point-to-point communications are necessary (e.g., visual sensing of moving objects in self-driving cars), the decisions will be pushed down to the lowest level. In other cases, cost considerations and the lack of a timeliness requirement may allow decisions to float all the way up to the cloud.
But, she noted, Dell is looking for commonalities among its different target customer sets so that the company’s IoT business can achieve some scale rather than having to create a custom solution for every situation. So, its cold-chain solution could work not just in manufacturing, but also in transportation, warehousing, and retail. Other promising areas include asset management (everybody has assets) and predictive maintenance (many companies have machines).
One of the hurdles facing Dell’s IQT effort is the nature of the enterprise buying decision. The person responsible for cold chain is not the same as the one responsible for IoT solutions. So, the pitch has to be made higher up the corporate ladder where the two converge and then brought down into operations in order to get everyone to cooperate for this new system they’ve never thought about, even if the business case looks great from the top level.
Difficulties aside, Dell see this opportunity as one of the richest in the future and is lining up to claim its share. At the event, the company revealed its intention to invest $1 billion in IQT R&D over the next three years and plenty more in operations. In my opinion, this effort is well conceived. IoT will be the largest growth area over the next five years, and it is still inchoate, amorphous, and coming together. With IQT, Dell has an opportunity to both form the market and participate in it, potentially taking a leadership position.