But what do people mean when they talk about cloud storage – and do they really know what it can do these days?
Not so long ago, once you got beyond Dropbox, for most professional users cloud storage meant online backup, and the biggest concern with that was getting your data back. That hasn’t gone away – some of the biggest cloud service providers make it effectively free to upload data but then charge you to get it back. So backing up your 5TB system is fine, but the cost of restoring it might sting a little…
Cloud storage is now a lot more than just that, though. Indeed, I’m constantly impressed by the new use models that people come up with for it. For instance, there’s now multiple ways to deploy cloud storage as a replacement for NAS. There’s even some interesting ways to deploy it on-prem in appliance form as primary storage, as for example Cloudian and Qumulo can do, targeting capacity-intensive and performance-intensive uses respectively.
Then there’s the ways that cloud storage can help with GDPR. Regulatory compliance means you need to do data governance and management properly, and that’s a lot easier if you have a single data store or central system of record. A consolidated backup and archive service is just that, and for some companies it’s easier to acquire and run in the cloud than on-prem.
Specialists such as Commvault, Mimecast and Veritas recognised this a while ago, Box and Dropbox know it too, and so of course do the major cloud platform providers. It’s clear to smaller players as well – for instance, when I met with Redstor recently, one of the opportunities we talked about was leveraging cloud-based data management for GDPR subject requests.
I see three things going on here. One is that hybrid cloud is now the norm, with the on-site element dealing with the latency and bandwidth issues that still afflict public cloud storage. Another is that although by its nature cloud storage is object-based, there’s more and more ways to deploy it as blocks or files to support existing applications – sure, born-in-the-cloud apps are object-based, but most current apps and end user requirements are file-oriented.
And then there’s our changing relationship with data, which is becoming virtual, like everything else. In this world, data lives “somewhere out there”, and a system backup is just a set of metadata that provides a view onto your data cloud that looks like a backup. A different set of metadata makes that data cloud look like a searchable archive, another view provides user self-service file retrieval, yet another supports GDPR data subject requests, and so on.
No longer do we need to think in terms of discrete systems, and of regular full system backups to tape volumes, real or virtual. Your governance systems still know where the data is really, of course, and apply the necessary security, but in data-centric ways, not system- or application-centric.
An over-simplification perhaps, but you get the picture. Today’s cloud storage isn’t what many people think it is – it might not even be in ‘the cloud’! In many ways, we’d do better to lose the term altogether and find a new one that better encapsulates those hybrid notions such as virtual-yet-consolidated, and local-yet-remote. If anyone reading this can think of one, please suggest it in the comments!