All Flash Arrays, Don’t Forget Your Helmet!

Super flash-blogIf you have a flash environment (Or know flash at all) you’re aware of its high performance. However, that high performance also demands the same level of speed in all aspects of its use. Meaning, you need to consider what you’re using to protect that all flash array, and know that you can’t skimp out on it. Not having the right data protection for your flash array would be like driving a racecar without a helmet – and everyone knows that’s a bad idea. Yet, the wrong protection solution can become a bottleneck to your mission critical workloads on flash, with your backup slowing down or even stopping your application.

When protecting a flash environment, there are some requirements to keep in mind. First, and foremost, is performance. Service Level Agreements are becoming stricter and your array needs the performance to meet these SLAs – which include backup windows, RPO and recovery time. Second is efficiency – flash storage, and the applications stored on them, can’t be impacted by protection. Having your flash array’s performance compromised is never an option. Last, but hardly the least, is agility. Since flash storage is usually a component of an ever evolving data center, it’s important that the data protection solution you implement evolve alongside your business.

There are three main pieces that are critical to consider when protecting an all flash array: Continuous availability, Replication for disaster recovery, and Backup and recovery.

Let’s start our discussion off with continuous availability, and how important it is to flash. All flash arrays give the best performance for the most demanding mission critical applications. These applications most often also need the highest level of availability. As the name suggests continuous availability is about keeping your mission critical applications always on.  Your availability solution should offer zero RTO and zero RPO even in the face of natural disasters and catastrophic hardware failures. In addition that same solution should offer a powerful data mobility engine that can eliminate planned downtime for time consuming storage tech refresh and workload balancing across arrays.

Replication for disaster recovery is something none of us want to think about – but have to. It’s like having a fire extinguisher, you don’t want to use it, but you definitely want it there if something happens. Your recovery solution should provide heterogeneous local and remote replication for operational and disaster recover, respectively. It should also provide continuous data protection, or CDP, for any point-in-time recovery.

Finally, let’s discuss backup and recovery. The information you keep on your all-flash array is critical, so your solution must be able to back it up quickly and efficiently. When discussing the type of backup that’s required for your flash array, you also need to talk about the protection storage that’s going to hold your backups. Reduction of storage and bandwidth requirements for your information is key – as well as your ability to replicate unique data from you backup for disaster recovery. Most importantly, you need to be able to reliably recover every time. But what about the backup process itself? That should share the same speed as well, shouldn’t it? You don’t want to – and you frankly can’t afford to – have your critical applications impacted by backup.

There are plenty of opportunities in life to be a risk-taker, but neglecting to protect your flash array isn’t one we’d recommend. Our advice is to find a partner like EMC that provides a continuum of data protection and availability solutions optimized for flash and designed to provide the performance, efficiency and agility required.

About the Author: Alyanna Ilyadis