Everyone keeps talking about the size of the Internet of Things (IoT), which will connect anywhere between 20 million and 38 million devices by the year 2020 — depending on who you believe. 

But size isn't the issue, according to Gartner's Paul O’Donovan. He claims the real challenge is building a business model to make it the IoT worthwhile.

That requires the development of IoT gateways — places were users connect to the IoT — as well as a way to optimize the volumes and varieties of data collected.

In a report published last week, O'Donovan noted, "We have entered a phase of intense competition for dominance of the IoT gateway market."

A Big Data Management Mess

Just consider the IoT in the home. The dawn of smart home appliances and ubiquitous sensors have fueled the dream of a holistic platform of residential devices and appliances. But it also gives rise to a whole host of security and data management issues. 

"Many IoT applications are triggered by sensors and need data management, but there is no single IoT gateway to the home," O'Donovan said. "As Internet-connected homes become increasingly smarter, the gateway becomes the center for connecting devices and appliances to enable the management of the ecosystem.”

He predicts a handful of vendors will develop gateways that control sets of sensors, applications or appliances, grouped through a single box or port.

It's a sizable market: The number of smart connected homes is expected to multiple from between 100 million and 200 million homes now to between 500 million and 700 million homes by 2020, he said.

The report urges businesses to recognize the potential of the IoT and develop strategies to manage and access to it.

IoT Business Models

As the IoT gateway market emerges, ISPs (Internet Service Providers) will be the early winners in the battle for the home gateway, provided they develop solutions or partner with hub manufacturers. Mobile phone providers will gain a smaller part of this market, but ultimately the cellular model will not have enough bandwidth to compete with the ISP solution, the report suggests

At the moment, however, the market remains fragmented and uncertain as everyone from vendors to regulators grapple with a host of issues ranging from infrastructure to access.

Technology and data vendors need to choose the information that passes through the IoT carefully. 

Part of the IoT business model involves the sale of data, O'Donovan said. "There is a huge issue here around data security and who owns the data," he said. While homes will not be sending a 24/7 stream of data over the Internet, it will stream data — and "deciding when is a core issue,” he said.

Think about it: All the devices in your home, some of them with several sensors, all sending information to a device provider.

O'Donovan cited the example of sensors for plants, designed to alert owners when a plant needs water. That information is conveyed to your smartphone while you are in work, for example. 

Data Overload

Do you really want that to happen? Probably not, if you evaluate the potential enterprise implications for massive data overload. 

“The amount of data will have to be restricted. Can you imagine the Internet being flooded with tiny little information packets from millions and millions of users about all kinds of things? We simply don’t have the infrastructure to manage all that,” he said.

Manufacturers building sensors for their products will have to program them to limit what those sensors send, just to make things manageable.

Sensors in cars, for example, are constantly culling information. But they should only alter the manufacturer with that data if it shows an anomaly or has gathered enough information to identify a trend.

But manufacturers are not data analysts. And to add this kind of data analytics into the manufacturing process would create enormous expense. This is where ISPs will come into their own. They will gather data from sensors and sell in to the manufacturers, O'Donovan said.