What is most valuable?
- Ease of use
- Performance guarantees that you can set
- Individual settings you can put on each individual volume, if you want to do that
- Ability to scale up, scale down whenever you want
- Scale-out ability, and the ease of adding a cluster - When you get a new node, if you farm out the datacenter tasks like we do, there is no technical ability required for them to plug it in and connect it and we can just add it.
- Open RESTful APIs are great
How has it helped my organization?
It's provided us the ability to not be concerned with setting SLOs for whatever application we're using. Everything is pretty much tier-one.
Our primary use case is virtualization, right now. We initially purchased it to be incorporated into our own internal cloud, OpenStack-based, KVM-based, so we use it for that. And, we've also branched into standard VMware as well. So we have both.
Based on those use cases we get really good efficiencies. We do a lot of encryption. We initially didn't have any because we were using it for anything, any LUNS, Oracle, whatever, and we didn't get the efficiency. So we positioned the use case over to virtualization and we're getting good efficiencies that way; to make it more cost effective. That's one of disadvantages, the actual cost. We haven't gotten there yet, but...
What needs improvement?
A little better segregation of the multi-tenancy. Right now, it's just VLAN-specific, that's all you can do. There's no authentication domain separations, things like that. For example, the NetApp product has storage virtual machines, which has a lot better segregation, and a lot better multi-tenancy, a lot better role-based access. That's probably the biggest thing that I would say, so we could actually use it for different tenants.
What do I think about the stability of the solution?
We've had some issues with hardware failures, and for them to resolve that it's actually meant replacing nodes. Otherwise, the stability is pretty good, all we've had is hardware failures. And they're built for a smaller scale, so before they were purchased by NetApp it was just, "replace the node, it's easier to replace." Any other field-replaceable unit, anything that breaks is, "replace the node," except the power supply or the disk.
What do I think about the scalability of the solution?
We're impressed with its scalability.
How is customer service and technical support?
For the most part, it's been pretty good. We've had to use it a lot for what I explained earlier, the hardware failures. This was mostly before it was integrated into NetApp support. We haven't really used them recently, so I don't know how the assimilation into NetApp support has taken effect.
What other advice do I have?
We purchased SolidFire for customer facing applications, they're all internal, house-developed applications that we sell to customers, to financial services.
When selecting a vendor to work with what's important to me are
- viability - are they going to be around?
Support is the big one. Is it just reactionary support, or proactive support? You need both of those.
I gave it a seven out of 10 based on what I've already explained. In the past, they seemed like more like a small company - and they were. But what I explained before, the hardware replacement, just replace a node. That's a small company.
Make sure that you have the performance requirement for it, because its price per gigabyte is a lot more than other solutions out there, if you don't need the performance requirements. You can get by on all-flash unless you have the need to guarantee performance on specific volumes.