What is our primary use case?
We're a large financial institution and we use them at our branch sites. They're definitely not enterprise-grade, but where we need to have an all-flash solution for applications, and a smaller footprint, that's where we use them.
The performance has been excellent.
How has it helped my organization?
The performance benefits weren't surprising, we expected that. What we didn't expect were the densification benefits that we got out of going all-flash. We're able to put more applications on the arrays because of how all-flash performs. The way some of the application profiles have responded to all-flash has been really pleasing. We got both a CapEx benefit and a benefit on the performance side.
Everything is sub-5ms for us. What I've found is that with all-flash, when an app from my business is slow, I pretty much know it's them and not me. It leads to a performance conversation that has really hit an interesting threshold point where we are better than what they need. So now we get to have that "refactoring your application" conversation a lot quicker because now the performance on the infrastructure side isn't in question anymore.
From an organizational perspective, the benefits are really around the decrease in performance-related issues. From an application perspective, when you think about incident management, those people are now spending their time very differently, where we have the SC Series deployed. It really has allowed us to advance a CapEx conversation, where our business partners, our LOBs, are actually behind us because they're seeing the benefit, real-time, in operations: the request for business cases to refresh environment instead of trying to sweat the current assets; now they're seeing the value prop. So we have a lot of people coming to the table supporting the expense, supporting that CapEx, perhaps a year or two sooner than they normally would have, because they understand.
What is most valuable?
Being a managing director I don't get "under the cover" as much as I used to, but what I can tell you from what my teams say is that it is a lot less intensive from a performance-troubleshooting perspective. The management software is a lot cleaner. We typically don't have performance issues anymore so our operations team gets to spend their time in different ways. We're getting to spend a lot more time thinking about how to deploy more efficiently and how to be creative about I/O profiles and tiering, in ways that we didn't get to do before, because we were troubleshooting performance issues, issues that we don't have now because of all-flash.
What needs improvement?
Overall, I'd like to see more synergy between Dell EMC's higher-tier platforms and their mid-tier platforms. What I have said, constantly, to my partners at Dell EMC is that the clear articulation of the path is really important to us.
In that vein, what I'd also like to see is, with the migration strategy that's built into this product, a lot more attention paid to Dell's - in particular - legacy platforms and how we get from some of our legacy EMC platforms onto this platform with a straight-through migration and scaling strategy, not host-based migrations and not piecemeal. What I've found is that when it comes to the "enablement wrapper," that's what I like to call it, or the "services wrapper," that is wrapped around this whole idea of a transition from legacy to new, Dell EMC would do well to focus more on my ability to skip a generation, rather than having me take individual hops because I can't greenfield my way into a software-defined data center fast enough. This journey of multiple hops is not helpful.
For how long have I used the solution?
One to three years.
What do I think about the stability of the solution?
So far, I haven't had an incident, so the stability is good. No downtime.
What do I think about the scalability of the solution?
It scales as well as the previous mid-range capability we had. I think for us, because we're in a situation where we're in flux between fiber channel and IP, I'm probably not the best person to relate to this question because a lot of our scalability issues have more to do with the network transformation that we're under and less so with the product.
We have in the neighborhood of 15,000 VMs.
Which solutions did we use previously?
We always invest. We're on a five-year depreciation cycle. At year-five, we do an evaluation of whether or not, from a refresh perspective, we need to move forward. The virtualization that's going on in our data center impacts that. In addition, we are green-fielding right now to software-defined storage, so some of these solutions are actually considered legacy right now. Applications are not refactoring at the same rate at which there is a desire to become software-defined, so we still have physical servers, some of which rely on fiber channel. Some of them, we can't even get to refactor into IP-based storage, even if it were still using a monolithic array. We have to provide our customers with what they need, so the need to invest has to do with meeting our customers' demands.
When looking at a vendor the most important criteria for me, specifically, are stability and reliability. Performance would be second, but stability and reliability are ultimately the most important thing. We have mission-critical applications and they just cannot fail.
What was our ROI?
To be honest with you, as a financial institution, I don't think we spend a lot of time thinking about ROI. We are so large, and our scale is so huge, that the economics that we're able to get out of Dell EMC when we go to the table, are quite remarkable.
For us, once you put a five-year depreciation cycle on top of that, we really don't have an ROI issue. What we're really about is ease of management, about automation, about understanding performance impacts, and that densification piece which really helps to pay the bill.
The one thing I can say, from an ROI perspective, is that the overall densification increase that we've seen, because of how flash seems to be operating with some of these application profiles, has definitely lead to a much better ROI. But it's not something we spend a lot of time thinking about.
What's my experience with pricing, setup cost, and licensing?
The partnership that we have with Dell EMC is great, so we always get to a solution that is affordable. We are A Dell EMC shop and we have been for a long time. A lot of that is not just because of the strength of the brand. A lot of that is because they build the best storage. For us, being enterprise-critical and having applications that process lots of transactions every day, we can't take chances on our applications and the ability of our storage to be reliant. Performance is important. Cost is absolutely important, but we're always willing to spend a little bit more for what we consider to be the best.
Which other solutions did I evaluate?
We're a multi-vendor shop. We do that on purpose to make sure that each vendor gives us their best. HPE was on the shortlist as well as some, what I consider to be up-and-comers, like Pure. They had a play but it wasn't a serious play. When it got down to the last two, HPE was the competition.
What other advice do I have?
Do your homework. Get in the lab. Figure out what it does and doesn't do. Figure out what's different. Understand your I/O profiles. Understand your applications. Think about how you want to best choose your app mix, when it comes to what's possible. Often, when people go all-flash for the first time, especially in the mid-range, they might be a little surprised at what's possible. Rack one up, fully populate it with disk, and really see what you can get out of it before you make assumptions around what you need and how your applications are going to behave.
We have not used the built-in capabilities for migration yet. What we're looking forward to is understanding the opportunity to possibly use these as a migration weigh-station between some of our older VMAXs or any of our other mid-tier storage platforms; where we might be able to use some of those migration solutions to help us get from legacy faster. We haven't done any migration between the two yet. All we've done is refresh tune-up.
I rate the solution an eight out of 10 because, overall, for the market that it's serving, it is just a really great product. For us, the extensibility that we've gotten in terms of being able to run multiple application workloads and still get a nice densification factor - and not have to worry a lot about over-provisioning and tiering or about a lot of the other things we used to have to worry about, thanks to the all-flash and the way that it operates - has been really nice. The management overhead is minimal. It just works. It's a workhorse.
Disclosure: I am a real user, and this review is based on my own experience and opinions.
Sep 09 2018