Perhaps lost in the current COVID-19 pandemic, is another pandemic of a far less threatening kind, the pandemic of the Internet of Things or IoT. It’s a good pandemic that will add value to everything we do and make. IoT is also one of the biggest drivers of storage demand currently exploding across the globe. It has been reported that there were 23 Billion IoT devices installed in 2018 with that number expected to grow to 75 billion units by 2025. While driverless cars are a good example of an internet-connected device with each vehicle generating terabytes of data every day, think about the smartphones that everyone has. Then think about smart everything else like smart homes, smart buildings, factories, machines, cities, airplanes, trains, trucks and so on. You start to get the picture and scope of IoT which is truly mind-boggling.
The market for IoT will continue to expand in 2020, especially as 5G networks start to proliferate. If you are like me and have been consuming more television than usual during this period of coronavirus lockdown, you may have noticed a lot of ads touting 5G networks. 5G promises to deliver 10X better performance than 4G, 100X better network density, with 100X more energy efficiency. With this key enabler rolling out, companies deploying IoT projects will need to plan for the data deluge that is coming as a result of these billions of devices!
To make these devices work better, or more efficiently, or safer in the case of transportation, the data they produce will need to be collected, analyzed, and stored for future analytics and reference. This is where Artificial intelligence (AI) tools come into the picture to provide the analytics power to derive value and competitive advantage from the massive volumes of data generated by the countless IoT devices. As a result, the combination of IoT data and AI will have a profound impact on the need for storage.
To be sure very high-performance storage will be needed to feed super-fast GPU processors used in AI applications. But AI applications such as machine learning require huge volumes of data scaling from terabytes to petabytes to train and test their algorithms. Organizations will then want to maintain access to original data sets for long periods of time and build on them to support a continuous cycle of data ingest, analytics, and inference. Extremely large archives will be needed as keeping all of this data on high performance storage will not be sustainable.
This is where intelligent storage management in the form of a more cost-effective and efficient active archive solution will be demanded to support the AI cycle.
An active archive is a cost-effective and always online scalable storage architecture that provides for rapid access to archival data via a virtual file system and moves data transparently between one or more storage platforms based on user defined policies. An active archive solution can leverage the unique performance and economic benefits of SSDs, HDDs, Tape, or the cloud (public or private). With additional intelligence provided by metadata and global namespaces, data silos are eliminated providing faster search and retrieval capability in a single virtualized storage pool with the ability to move much more data than legacy solutions typically challenged by moving massive amounts of data.
As volumes of data increase from IoT devices and to feed AI applications, cost and performance will play a larger role. What’s needed are storage solutions like active archive that deliver on these requirements. Look for active archives to be an even bigger buzz word in the storage industry, maybe even the next IT pandemic!