Do you even need on-premises backup anymore?
It’s a valid question. After all, use of the cloud as part of a backup and recovery strategy comes with the reliability and speed of modern Internet connections, the high availability of cloud storage and compute, and the associated security needed to protect your customer’s most important data, applications, and systems. All this makes backing up to – and recovering from – the cloud an extremely viable option for your customers.
So, in a world where the cloud can actually deliver on what it promises, should everything go to the cloud?
Before I jump into this, please keep in mind, this is most definitely not a “bashing the cloud” blog – it’s quite the contrary. The cloud has moved backup and recovery forward like no technology and is not just a viable option for backup and recovery, but is a necessity. So this blog is more about, perhaps, going a little too much to the cloud, and thinking on-premises isn’t even needed anymore.
It’s tempting to simply go all-in with the cloud as your backup repository and recovery target. But it doesn’t necessarily mean leaving on-premises completely out of the picture is a good idea for every one of your customers. Here are a few reasons why:
- Some “disasters” don’t require the cloud – we have thrown around the term Disaster Recovery, implying every recovery event is a disaster. But some events require little more than restoring a file or a database to a local server. While the cloud can provide the value of, say, unlimited storage (compared to an on-premises storage array), none of that necessarily comes into the picture when you consider the specific recovery needs of a given “disaster”.
- Local storage can be faster – for the record, this isn’t always the case, but depending on what’s being recovered, pulling from a local repository can result in a faster response time.
- The cloud is not a guarantee – Even the cloud is susceptible to outages. After all, there’s a lot that stands between your customer’s network and the cloud storage you use: your customer’s internet pipe; DNS; routers; cloud front-end servers for authentication and encryption; back-end servers for storage and compute; etc. If any one of those has issues, it’s possible that the cloud may not be available when you need it.
- Some systems can’t (or shouldn’t) be recovered to the cloud – Some systems, such as high-end databases, won’t run well on virtualized hardware and need to run on specifically configured physical systems, making the need to recover locally a more viable option. In other cases, the recovery of, say, the CEO’s laptop to the cloud just doesn’t make sense.
- Local is a viable recovery target – Even when you need to recover a VM quickly, most backup and recovery solutions provide some type of “instant virtualization”, allowing the recovery and spinning up of a VM locally without the need to invest in a virtual infrastructure.
The reality is that in order to provide recovery of every data set, system, and application, in the face of each possible disaster from which you wish to protect your customers, you’re going to need a hybrid mix of on-premises and cloud-based storage and recovery options.
Putting the cloud up on a high pedestal is probably warranted with the capabilities available today in backup and recovery. But don’t count out on-premises; it has a place in your protection strategy, ensuring your customers always have the best coverage possible.