There are some things that simply have to happen, no matter how modern your business and your IT systems. One of these is obviously ensuring that you have effective backup copies of your business data available.
Yet even though this topic is arguably more important than ever, it is one that is easy to overlook, or at least to not pay enough attention to. Until something goes wrong, that is, whether it’s a system problem, data corruption or simply a user doing something silly. In this respect, data protection is a lot like insurance: no one ever wants to pay for it, but they are very glad they did when they need to make a serious claim.
So, given the importance of backup and recovery, a question for today is should we still be backing up to tape and disk, or is it time to use modern, fast media? It is clear that All-Flash Array (AFA) systems are well suited to a wide range of workloads. And as the price differential between AFAs and hard disk arrays continues to narrow and AFAs become more affordable, that range of suitable workloads is rapidly expanding.
Why do backup and recovery need such performance, though? The main reason for moving data protection onto flash is the same as it was when hard disk systems began to displace tape as the target for backups: speed. It’s not just performing backups faster though. Every IT Professional knows that business users’ expectations for availability mean that applications and services must be backed up ever more frequently, perhaps several times a day, perhaps many times an hour. This makes backup windows tiny and means that the physical copy of the data to the backup platform must be fast, sometimes very, very fast.
More important still is the speed with which data must be restored. “Time costs money” is a very old saying, but in today’s business world it is also true to say that downtime costs money, and the sums involved can on certain occasions be matters of board-level concern. The talk is often about the opportunity cost of downtime, never mind the impatience of internal users and external customers.
Waiting days, or even hours, for data or systems to be recovered is no longer an option for any business-critical service, therefore, and is practicable for fewer and fewer mainstream business systems and user data.
What about NVMe, then?
So if recovery does require better – and sometimes the best – performance, does NVMe also come into play? After all, the high-speed NVMe and NVMe-oF protocols were designed to allow all storage media, but solid-state drives in particular, to operate with the lowest possible latency. And with data being held in widely dispersed locations, NVMe-oF could also be called on to extend traditional data protection solutions beyond the data center. This has the potential to dramatically increase their business value.
As the value of data protection is established, it is fair to ask if the cost premium of NVMe solutions is justified by the additional benefits? NVMe and NVMe-oF solutions still carry a cost premium, but given that the value of data protection is well established, is that premium justified by the additional benefits they can bring? At some time in the near future, it almost certainly will be. Indeed, a recent survey by Freeform Dynamics says that for a small number of organizations, that time may have already arrived.
The importance of backup and recovery is well understood by IT professionals, but getting business buy-in to do it properly has always been a problem. However, as shared storage becomes pervasive it may be time to integrate the recovery infrastructure into the bigger picture.
Treating backup and recovery as isolated from the core IT infrastructure worked in the past, at least to a degree, but it makes much less sense today. Speed and resilience are of paramount importance, so backup and recovery require the same capabilities.