The first cloud wave, with test environments or less business-critical services has passed. Nowadays, business systems and important industrial services are moving out into the cloud. This is making our industry more efficient.
Locally collected data on the health of a robot at a car factory has a low value. But once this information is instead aggregated with other robots in the chain, the manufacturer's own data and the owner's production plans, the information suddenly has a new value.
“We want to take the data from machines and make predictive analysis: When does the machine need servicing, what spares are required, does the production line need new robots, and what about insurance?” says Tarmo Pajunen, a Sustainability Ambassador and IoT manager at Cybercom.
Gartner predicts that by 2018, six billion connected machines and systems will require ongoing maintenance. And by 2020, thirty-five billion things will be connected to the internet. As a result, companies will be forced to look at machines and systems like customers, and even treat them like customers. The data generated will be plucked up from each machine and processed centrally. In the cloud.
“For this to work the data must be stored centrally, where there are databases, computing power and and application program interfaces (APIs) – and of course everything needs to be scalable, so we see a future of machine dialogue in the cloud,” says Tarmo.
Tarmo is one of the people behind the Machinebook concept. By sharing and reacting to information (imagine Facebook’s thumb), the system learns who is interested in, for example, an alarm from a machine. The aim is for information to move from dashboards in a control room to the right operator tasked day-to-day with making sure that production lines and large machinery are working properly. This makes the company less dependent on the skills and availability of an individual operator, and instead builds up shared expertise.
“In the future we hope that machines will be able to learn from people by listening to human conversations,” says Tarmo.
He argues that machines with locally stored data quickly become a problem since storing the information locally means the common information becomes incomplete if a machine loses contact with the network. And building up the cognitive intelligence to be able to understand what people are saying requires computing power that cannot be housed in an individual machine.
“You don’t want to have the intelligence physically attached to the machine. It also uses unnecessary electrical power and local computer space.”
According to a survey of 770 major European and North American companies conducted by Forrester Research, more companies are currently evaluating a transfer of their business systems to the cloud. In two years, interest in the cloud grew from 12 to 35%. Tarmo believes there is now a similar shift in IoT. This is because the options for local in-house operation are often too expensive and do not perform adequately to provide any benefit from the collective intelligence of all the connected machines. One consequence is that IBM has increased the pace of IoT functionality in its public cloud.
“Connected machines themselves are becoming more the norm. The real benefit occurs when this is done in a smart way, thinking about how to improve processes or save money by using all this data,” says Tarmo.