If you believe the self fact-check, the annual Gartner strategic predictions are right 78% of the time. This might make the three-to-five year tech outlook worth paying attention. Here is a look at a few of the 10 predictions.

Through 2019, every $1 that enterprises invest in innovation will require an additional $7 in core execution.

Innovation isn’t cheap. It’s hard to be innovative in a status quo environment, but innovation in a vacuum is equally as likely to fail. Gartner advocates a “bimodal” approach to managing IT. Mode 1 is based on predictable improvements in well-understood areas. Mode 2 is exploratory. It looks to solve problems and create breakthrough ideas.

One area that Gartner sees investing heavily in innovation is IT service providers. They expect that by 2018, ISPs modernization work to support digital innovation deployments will increase by at least 30% over today’s spending levels. At Uplogix, we work with a number of ISPs in markets ranging from enterprise to satellite connectivity services and we hear of a number of initiatives aimed at modernization.

The Uplogix platform helps ISPs reduce costs of deploying and managing network infrastructure while improving service levels, providing bottom line savings that free up cash and valuable human technical time to work on innovation programs.

Through 2020, the Internet of Things will increase data center storage demand by less than 3%

This one is kind of counter-intuitive. It’s been part of the fear of the rise of the IoT – explosive data growth as every sensor reports minute changes in temperature, location, health… you name it — gobs of data clogging the internet and data centers. Garter says most of the data generated via IoT sensors and devices won’t be stored.

They break IoT data of 2020 into consumer and business applications, with consumer applications accounting for only about 3% of the data and business IoT applications generating 97%. The key to managing storage is to not assume that all data has to be stored. Filtering out steady-state trends, developing a data reduction strategy based on regulatory requirements and industry best practices will go a long way. Then there is the simple check — is the projected economic value to keeping the data worth the cost of maintaining it?

This is kind of similar to the approach Uplogix uses for network monitoring. We gather tons of device data and service level statistics at the local level. When things are nominal, there isn’t reason to push all of that data upstream. Keep data that is useful for troubleshooting issues and report on the unexpected. Otherwise, assume that everything is OK.

By 2022, IoT will save consumers and businesses $1 trillion a year.

According to Gartner:

“Assets under maintenance globally exceed $240 trillion, with maintenance costs of $27 trillion. Moving to predictive maintenance can often save 10% to 20% over preventive maintenance. IoT, properly deployed, can unlock this $1 trillion savings opportunity.”

Gartner predicts two extremes for IoT projects designed to reduce the cost of maintenance. The first is more robust, inexpensive monitoring delivering analytical data that can be used to spot patterns and predict maintenance needs. The other extreme is the idea of a “digital twin.” It’s a computer model of real world component or process. Real-world data from the physical twin is applied to the digital twin model, along with additional data like weather and analysis. The digital model then provides guidance for management of the physical twin.

Uplogix is closer to the robust, inexpensive monitoring model. By removing the network from network management, we’re able to continuously monitor devices and detect state as well as track trends. Where we go beyond the IoT model is that we can respond to issues with actions, automatically, and faster and more reliably than centralized management solutions.

Look here the other Gartner strategic predictions for 2017.