IBM estimates that smartphones and devices related to the Internet of Things will generate 44 zettabytes (exabytes to the thousandths) by 2020. To handle all that information, IBM is looking at storage solutions that combine machine learning and artificial intelligence, both of which excel at finding patterns. Other companies are looking for solutions, including Pure, which is combining flash storage and engineering to create a storage unit that currently holds 16 petabytes, or five times most storage devices.

The New York Times reports that Pure co-founder John Hayes says, by 2017, his storage unit will hold twice as much, using just 4 percent of the current power consumption. Dell bought storage industry leader EMC, and Netflix “ships dozens of custom-built appliances a week to hundreds of local Internet service providers,” a mix of flash and traditional storage holding five petabytes.

Netflix executive Ken Florance notes that the company plans to offer “10 times the performance of competing products, at a lower cost.” “This market is at least as large as our original storage business, and a lot bigger over time,” he said. “I’ve talked to companies planning for hundreds of petabytes. Even a body camera on a cop adds up data pretty quickly.”

The Wall Street Journal reports the details of an International Data Corp. survey showing that 58 percent of companies plan to use “Web-based, on-demand computing services” for more than two applications, up from 24 percent 14 months earlier.

In response, Amazon and Microsoft and others are seeking new revenues by easing that transition. Amazon is launching the AWS Database Migration Service and Microsoft said SQL Server 2016, its next version of the database program, would include technology to help customers transfer data to Microsoft’s Azure cloud service. Oracle has also said it would help customers migrate to its cloud services.

Moving all that data is an arduous process, says Gartner analyst Lydia Leong, who reports that companies with $100+ million annual revenue can take up to two years to migrate their databases. AWS vice president Adam Selipsky reports that more than 1,000 databases from “many hundreds” of companies have availed themselves of an early version of Amazon’s Database Migration Service.

Some companies are moving in the opposite direction, taking their databases off AWS and creating their own networks. That’s what Dropbox did recently. According to Wired, “over the last two-and-a-half years, Dropbox built its own vast computer network and shifted its service onto a new breed of machines designed by its own engineers,” with software written in a new programming language and built by its own programmers.

Citing Forrester Research’s data that, by 2020, cloud computing will be a $191 billion market, Wired emphasizes just how important the cloud is becoming. Although many will rely on third party cloud providers, the bigger players, says Dropbox chief operating officer Dennis Woodside, might get “substantial economic value” by creating their own solutions.