An Evolutionary Journey of Enterprise Storage for an SMB: Part 3
Welcome back to part three of our blog series, discussing how we came to our decision about which storage solution was best for Evolution Recruitment Solutions. If you missed part one and part two, be sure to give them a read before you find out about the storage solution we decided on!
We finally decided upon a Dell Compellent SC4020; it looked attractive, as it was All-Flash with a good balance of capacity against cost, as well as five years’ warranty and Copilot support. The Compellent range offers storage replication to EqualLogic at no extra cost, so there’s potential to reuse the EqualLogic for DR.
Dell was not only offering us a great deal for the Compellent storage appliance, but they also offered two N4000 10GB switches to replace our existing 1GB fabric switching stack, 10GB NICs for every host, four N2000 PoE series switches to replace our desktop stack, and all this could be wrapped into a three-year 0% finance deal, which our finance department loved!
Comparing their products against their competitors, their features and performance were good and at a reasonable price. Even though I didn’t feel it was the best storage solution, or that their network switches were the best we reviewed, sometimes these aren’t always the reasons why a decision is made. In the end, this mostly came down to a commercial decision; not what was the best.
The deployment was fairly straight forward, but we did have assistance - all Compellents must be installed by a certified Compellent engineer, in order to validate the support warranty with Dell’s Copilot support. For the migration, we simply added the Compellent to our vSphere environment and vMotion’s VMs. Following the deployment of new switches and storage array, the increase in performance was huge across the board, and with the additional capacity which had been a blocker to a lot of projects in the past, finally the flood gates were open.
In fact, this lead to further issues; within eight months we had filled the Compellent and needed to increase the capacity. As we had spare drive bays, we requested a quote to install an additional three 1.92TB SSDs - we got a shock when we then discovered that we could only purchase a minimum of six disks, which cost around £10,000! Dell had specifically told us that we could just install a single extra disk if we wanted; this was a key point we highlighted for review when comparing the solutions.
I called Copilot support, who confirmed that, to their knowledge, you actually can just install and expand the capacity by a single disk. There was nothing stopping it happening from a technical standpoint; I have even been to Dell EMC events where this has been stated numerous times in their sales pitches. So why in reality was this not the case? A key lesson learnt - don’t believe the sales pitches, they often bend the truth, and sometimes they don’t even know they are doing it. The left hand and the right hand don’t always know what each other are doing, and this isn’t just a problem at Dell; I have come across this a number of times over the years, and sometimes there isn’t any way to account for it.
You can try to avoid this happening by asking all the right questions, but if the answers you’re getting are simply, unwittingly wrong, then it’s sometimes unavoidable. It’s always really important to make sure you are comparing apples with apples; sales pitches are always full of bold claims. Every vendor claims they have the best compression, or the best deduplication, or the highest maximum number of IOPS, and some don’t even measure in IOPS so they can’t give you those metrics. Often it felt like looking through smoke and mirrors; comparing all these storage solutions proved to be quite difficult.
Our decision could, in the end, prove to be short-sighted, as we may have future projects surrounding backup and DR which will require server resource. This would probably mean convergence might have been a better fit. We could just let our servers’ warranty expire all at the same time, which would enable us to refresh an entire cluster at the same time with a converged solution. Evolution’s datacentre is on a journey; as technology changes, so does the datacentre. We can only try to make the best decisions that are correct at that point in time.