Annual IoT data collection to reach 1.6 zetabytes by 2020, says ABI
April 16, 2015
The numbers are beginning to look silly. ABI Research is estimating that the annual volume of data captured by IoT-connected devices will by 2020 surpass 1.6 zettabytes, that’s 1600 exabytes, or eight times the 200 exabytes it estimates were collected in 2014. An exabyte is a billion gigabytes.
“The data originating from connected products and processes follow a certain journey of magnitudes,” said principal analyst Aapo Markkanen. “The yearly volumes that are generated within endpoints are counted in yottabytes (1000 zettabytes), but only a tiny fraction of this vast data mass is actually being captured for storage or further analysis. And of the captured volume, on average over 90% is stored or processed locally without a cloud element, even though this ratio can vary greatly by application segment. So far, the locally dealt data have typically been largely inaccessible for analytics, but that is now starting to change.”
In terms of deployment architectures, the IoT is undergoing a major paradigm shift from cloud computing towards edge computing. On one hand, this shift is opening up edge-based data to meaningful analysis, by distributing the analytic workloads across the network. On the other hand, it is also shoring up the cloud-level capabilities by making the transmitted data more actionable, by enriching and contextualising the payloads.
“Edge computing is a huge challenge for the entire IoT value chain, as we can see from the way that cloud platforms, analytics vendors and gateway suppliers are scrambling to collaborate with each other,” said ABI practice director Dan Shey. “It is also a great opportunity for various software and hardware players that have been working towards this goal even far before the IoT as a concept became fashionable. Names like AGT International, Eurotech, Kepware Technologies, OSIsoft, and Panduit are all examples of firms whose background in distributed intelligence allows them now to expand their target market, even significantly.”