NetApp AFF - All Flash Reviews

It reduced the overall latency in our Citrix infrastructure. We have a pretty robust Citrix infrastructure. Before putting in the All Flash FAS, end users would see a lot of latency. That's been the biggest improvement, along with a lot of improvement in the overall performance of our SAN and a few other data intensive applications.

There was noticeable latency before the All Flash FAS. Since the All Flash FAS, it is extremely fast, no latency whatsoever.

I would basically just like to see improvements with the reporting; consolidating metrics, performance and any sort of issues. Right now, there are a lot of different tools, a lot of different places to go to see the overall health of the system. I would like one place, a dashboard, to see everything. I know there are some things that NetApp has released and are releasing, but we haven't gotten to the point where we've implemented those yet.

It is extremely stable; never had any down time or issues with it. It's fully redundant. All of the updates have pretty much been non-interruptive; it’s an extremely stable platform.

It scales out well. It’s a new All Flash FAS and we looked at the overall capacities that we needed before. It's only been in place for about six months. From a scalability perspective, we know that it will scale out if we need it to, but it's a new implementation, so no issues or anything like that.

So far, we haven't needed to use technical support for this one, yet.

We did not previously use a different solution. We were a NetApp shop before that, but we were using a different controller and we weren't an All Flash FAS shop. We could see the latency. We used all the utilities, so we could see what was going on, the need and how it would help our business.

Initial setup is generally straightforward, but NetApp has good technical articles and guidance on moving from one NetApp controller to another NetApp controller. It was pretty straightforward for the most part.

We looked at a number of other flash systems and solutions for our latency issues. At the end of the day, we just decided to continue and move forward with another NetApp controller.

Reliability and availability are the most important criteria for me when selecting a vendor to work with. We need them to be available. There are a lot of vendors out there that have a lot of people, but if you're building a reputation and you can't get the people you need, then it's a problem, regardless of how good the controller is.

It does what it’s meant to do; works extremely well in our environment. We have multiple data centers and the replication works really well. Overall, it's pretty easy to use.

Look at your individual company's needs. In general, look at your nice-to-haves and must-haves, and then weigh the options and see what works best. NetApp has been a great, established company. We've had a good relationship with NetApp for a long time and so we would recommend them to a colleague.

The most valuable feature is the IOPS speed, everything that comes with it. It's just a great platform to be on, for example, with VMDK files, data stores, and VMs in general. Users say things like, "What happened? How come it's so fast. What did you do different?"

All the features that we were sold and told about, they all work; it's been good.

First of all, the cost was a benefit to my organization. The cost was great for us. It just made sense to do it. Then, speed. Then, just overall manageability of the system itself.

Higher performance would be an improvement, absolutely.

They could bring the cost down but, as I’ve mentioned, the cost was right for what we needed.

Regarding whether you are happy with the user interface, I think that depends on whether you're used to the CLI or you're used to the GUI. I'm a CLI guy, so I like it.

I've had zero stability or scalability issues.

For the AFF, I haven't had to use technical support; I'm good.

We were on a very old 7-mode system; that's what we migrated from. That was our next step, to stay with ONTAP because we liked the features of ONTAP, and we wanted the speed and performance of the All Flash FAS.

Initial setup was very straightforward, just simple. It wasn't difficult at all.

Follow the instructions. That's all. It's straightforward; it wasn't hard at all.

Pricing was very competitive but right on the mark for us.

Before choosing this product, I also evaluated EMC and Pure. The decision came down to what we were used to managing and what we trusted.

In general, when I’m looking at a vendor to work with, I look for honesty. That's all I look for. I understand they have to make money and I understand we have to spend money to get it, but I don't want to be taken; I don't want to feel like I'm getting taken as it's being sold to me.

The most valuable feature is the performance that we got out of it. With a previous solution, we had some latency issues and performance issues. When we got the FAS All Flash Array, that technology took care of those issues that we had, those bottlenecks.

It provides greater stability for our corporate database, which we host on the FAS. We have a much greater sense of confidence and reliability in our data solutions. It gives us more confidence that everything's going to keep working.

In terms of manpower or cost, because we are a public agency, it's more about value as far as the service that we receive and the stability of the solution. Those are the key factors.

I'm not sure about room for improvement, only because right now we've just completed a major upgrade. At this point, we're very happy. We don't see anything lacking in that regard.

Nonetheless, there used to be a product called Balance and it's been really replaced by something called Insight. From an operational perspective, the ease of use, we preferred Balance. Even though that product has come to end of life, we're unhappy about that.

OnCommand Balance and Insight are two separate NetApp products, that provide the performance capture and logging features. OnCommand Balance is an older product and NetApp has announced that it will not continue this product anymore, as its replacement is Insight. The staff here have used both of these products and they prefer OnCommand Balance. Sadly, we won’t be able to continue using OnCommand Balance, as long as we would have liked to.

So far, we've had no issues whatsoever with the stability. It's beating our expectations for an enterprise-wide solution, whereas other solutions that have presented themselves as enterprise solutions haven't performed to the same degree.

We've never had any issues with NetApp. In particular, the customer service I think has been far superior. Our business decision was basically based on NetApp's record with us for their customer service. We're making NetApp our single storage standard within our organization.

We previously used Oracle. There was a Pillar Axiom line for storage. We also previously had an EMC solution. I don't remember exactly what line that was.

We decided to invest in a new storage solution just because of the data growth that we needed. We're expanding our business content, meaning disaster recovery architecture. We needed to expand to an additional site.

As I’ve mentioned, we had Oracle’s Pillar Axiom line. We also looked at Compellent, which is Dell now, and Dell fired them. Then HP. We used to have an HP EVA as well. We used that before. We looked at HP's current solution. We weren't happy with that one.

We decided to go with NetApp over HP because of the experience we had with both of those organizations in customer service. NetApp, again, was far superior. Our requirements then to our reseller, or VAR, and NetApp was that we knew what our workload was and we needed to have a solution that would meet certain criteria, which was set on latency and bandwidth thresholds. The vendor, along with NetApp, was able to provide us with an evaluation unit that met those specs with flying colors.

Understand what your workload is first. What is it that you're trying to accomplish so you set the proper thresholds and criteria for performance. Understand what your support service needs are. Is that important? How important? It's not always about cost. We found that in all those areas, with our evaluation, NetApp was a clear choice for us, based upon past experience. We continue to have success with NetApp.

Multi-platform support is one of the most valuable features. It has lots of data protection solutions and cool new features, such as vol moves and FlexCloning. That's very useful for database refreshes.

We heavily leverage the FlexClone features to clone databases for various environments. We use the multiple protocol feature to support different operating systems and platforms.

It allows us to be more flexible with customer demands and needs. It has not allowed us to save money, per se; there are other solutions that are probably cheaper in the flash arena, but this was a nice transition from our NetApp 7-mode to CDOT platforms.

I’d like better performance management tools and a federated provisioning tool to manage our storage. They're working on that right now. They don't have anything out of the box that comes with that at this time that I know of.

I've been using AFF Flash for about eight months; we tested it 12 months ago.

It's a very stable platform; so far, it’s been very stable.

It scales very well. I'd give it about a 9 out of 10 on scalability.

They have strong technical support. We've had some issues in the beginning with the technical support because it was a fairly new product, but they seem to be scaling up in terms of their support engineers.

We evaluated Pure and Tintri. We're an incumbent customer of NetApp’s 7-mode product, so for the migration from 7-mode to CDOT AFF was easier than transitioning to Pure or Tintri.

Some of the competitors did not offer multi-protocol solutions, so the architecture for those solutions would have been a little bit more complicated.

Me and my team did the initial setup. Setup was more complicated than their 7-mode platforms, but it's a necessary evil to provide the new functionalities within CDOT and AFF.

Moving to a flash solution was definitely beneficial.

If you're not already in flash, you should probably start thinking about just buying just flash. Flash helps relieve some of the performance capacity management overhead that comes with traditional spinning disk platforms.

What I would suggest to people that are looking at flash is to make sure they're able to do proper sizing. With buying flash, you need to also make sure your controllers are able to support the workloads you expect the flash to handle. I think flash removes the disk as the bottleneck, but then that pushes that bottleneck down to other hardware components, such as either the network SAN or storage controllers. Make sure that the rest of your system can handle it.

That's what I would offer in terms of evaluating a flash solution, and to look into scaling out versus scaling up for flash.

It is cost-effective flash for us. It's a platform that we've used for quite a few years. We've been NetApp customers for probably about eight years right now. You don't have to go in and re-learn any kind of new interface; it's using basically the same interface. Provisioning is ultra-fast and it just works.

We had a few databases that were using SQL databases that we were having some performance issues with. We moved them over to FAS and, no more performance issues. Basically, you throw a ton of hardware at a problem and that fixes it.

A lot of the applications that we use are canned applications. We don't actually have the ability to go in and modify them. We’re kind of handed a bad deal in some aspects. We go in, we put that stuff on flash to see whether we can make this thing perform the way it's supposed to. We really don't have the option of going in and changing the code.

I recently attended a conference and one of the sessions was about performance data with ONTAP 9. They've addressed some of the issues that I'd like to see in that, such as being able to see where your latency is, and how much performance you have left in the array before you need to start looking at what we need to start moving workloads around.

It would be a little bit nice if the monitoring was a little bit better and smoother, but we've not had any issues from that perspective. In the future, I don't want to have any of those issues.

We've had no stability issues so far with it; in fact, I've not lost a disk out of it.

Scalability is not something that we're going to be concerned with right now, as far as adding; we can always add a tray. It's non-disruptive. That's great.

For flash, we did use professional services to come in and help us get it set up but other than that, we've not had to make any phone calls about it. It's pretty straightforward.

At the time, we had been NetApp customers for quite some time. We had been using a FAS3220 and we were starting to see performance issues. Our sales engineer said, “Why don't you guys try take a look at this?” We did some research on it. We actually POC'd it with a few others, that will probably remain nameless at this point in time, and obviously, NetApp outperformed the others; oh, we loved that.

Well we were a 7-mode shop and we were switching to CDOT. There's a little bit of a transition there. I won't say it's overly complicated; it's just some new things to get used to.

The setup complexity is why I have not given it a perfect rating; not that it's a big deal. We had professional services come and help us do that but going from 7-mode to CDOT was quite a jump.

I have a feeling that’s pretty common. We were going through the conversion on all of our arrays. We currently have three. It's getting easier as we go through the process more and I understand it. It would be a lot better if the transition was a lot more smoother.

A lot of companies will tell you that they're the best at what they do. As a company, I think it's very important that you look at POCs to see if you can get them. Everybody can tell you they have the best product, but until you can actually prove it on your workload, you really don't know 100% for sure.

When selecting a vendor to work with, as a company, we have had a tendency to just go and buy the "best of breed," which sometimes included arrays from multiple vendors. As a company, we have five different brands of arrays. You can't become an expert in something if you have five different arrays to work with. What we're trying to do as a company is to align to say, best of breed being, this is fantastic as a NAS appliance so we're going to look at that and say, maybe we should look towards getting that. I think we're taking our shotgun approach and we're kind of moving it down to where you can be more specialized in what you do. As I’ve mentioned, NetApp is fantastic; it does block, it does NAS. It's a one-stop shop.

Performance is the number-one most valuable feature, for sure. Flexibility and the multi-tenancy are also valuable.

The compression we needed, the rates we get, are inline with the performance, which is the reason we bought it; we have a lot of applications that use it. The compression and the dedupe stays in storage but on our other products, we'd lose performance because of that. On the All Flash FAS, we don't have any performance issues at all, so it's a big differentiator for us.

It provides financial benefits, because we don't have to spend as much on storage, because of the dedup and the compression and the performance it gives us. We don't have to buy anything else because of it.

There’s one thing that would make it easier to work with. There are differences between using the OnCommand: the GUI vs command line. There are still differences. There are things you can do from the command line that you can't do from the GUI. If they could make the GUI do everything that the command line does, that would be the best. That would earn it a perfect rating, for sure.

There are certain configurations/settings on cDOT that you can only make by using the CLI. My point for room for improvement was that, if they could make all the configurations/settings available in the GUI, then you would be able to pick one or the other for managing the cluster. Today, you either have to only use CLI or a mix of GUI/CLI.

We've had no stability issues. We've never had a problem. We've only been using it for about six months, but we haven't had a single issue of any kind. We're happy with it.

We've added on shelves to it. That's one of the reasons we bought it too. We bought it for a certain set of applications and we've already expanded that now; used it for other things too. That's why I bought more storage on it. The flexibility we have, all the connections it has, it's helped us without having to buy either more storage systems or other products. We've just been able to grow what we have.

We've previously had several other vendors. We used Hitachi. We used their HNAS product. We had Celerra from EMC. We've had a couple of other older vendors that aren't even around anymore.

We switched from HNAS because of the performance, both in application and backup performance. It was not nearly what it needs to be. Their storage pools and the way we could grow the HNAS environment was nothing compared to what the NetApp does. All of those things together made that an easy switch.

It's definitely saved us in storage costs. It's saved us in reliability, in downtime. We’ve had downtime with our HNAS, a couple times. That was the factor that got rid of it in the end. We invested in that product, and it was a pretty important feature of some of the applications that used it. We kept going with it and staying with it because we invested in it. But we had too many outages, too many problems with it.

In the end, we decided that it was not worth it, financially, to keep it. We got rid of it, and invested in NetApp, and all those reliability and performance issues went away. It's been 100% since day one.

We've had other vendors, and we've used their solutions. The performance hasn't been what it is on NetApp or the compression dedup rate hasn't been what it is on NetApp; with those other vendors, we get one of the two. We get both of those with NetApp; better performance, better compression, all of those things without sacrificing performance.

Look at NetApp first. The flexibility they offer, the performance, and all the features they have. I can't think of anything that we can't do with that product. That's where we go to first now. We have a lot of other products. We have a lot of other storage vendors: Hitachi, IBM, EMC. We've had other NetApp FAS products, not just the All Flash one. We still have other NetApp FAS products.

Since we've had the All Flash FAS, because of its reliability and everything that goes with it, it’s the first thing that the application people ask for. When we talk to them about needing more storage, they always ask for NetApp first. It's kind of the standard now, which is fine by me because I like it.

It's reliable; it's fast; it does everything that we need it to do; it's relatively easy to work with.

When I’m looking at a vendor to work with, I'll be blunt: I don't want to worry about it. I don't have to deal with it. I don't want to work with it a whole lot. I'd like it to be, "I have a use for it and I want it to do that". I don't want to have to do a lot of configuration, tweaking or perpetual updates and patches. NetApp, specifically, was very, very good about one-time patches, no-downtime patches. They work well from a it-just-works standpoint. I can't always say that about all my vendors.

There is room for improvement in the migration step. The no-copy copy has a lot of caveats. We felt like they weren't brought up early in the process. They were gotchas as we moved through the process. None of them were earth shattering or show stopping, but they often resulted in another night of work, another evening worth of work, or we had to shut down this over the weekend to make this happen, that we weren’t really planning on. That would be the change, but that's just at this point.

These setup hangups are why I’m not giving it a perfect rating. It’s close to being perfect, but I am going to have to round down in this situation.

For more detail:

We had a head fail – fortunately everything failed over as expected, but obviously having a system that’s less than 4 months old fail is hard.

Post hardware repair the new head needed a manual intervention to get all of its firmware where it needed to be, which did cause a downtime for CIFS users. We were able to schedule for after hours, but downtime is not something we ever really want in our storage systems.

So far, it is a stable solution. We haven’t had it that particularly long. We haven't had many issues with it.

It’s met our scaling needs.

Data is growing everywhere, so we’ll work with the data we’ve picked up and it will help us for the next calendar year. I fully expect we'll need to add more more shelves within a year.

Technical support is excellent. They are very responsive to our questions. As I’ve mentioned, we did a 7-mode cluster transition; there was a lot of learning on our side and they have been very patient with keeping us informed, getting us up to date on what we need to do on the storage side end.

We're still in the VMware environment but we moved from an old 7-mode cluster using old SAS drives to a new all-flash pool, mostly for performance purposes. We had some expectations and we're blown away. They all paid off, to the point the users actually saw big changes as well. We knew we'd see things on the back end, but we weren’t sure they would trickle down to the end user.

Initial setup was both straightforward and complex. We used their tool that provides a no-file-copy copy. The goal was to switch all of our existing shelves to our totally new heads. It was sold as being fairly painless, no-change process. In practice, there were a fair number of changes, a fair number of tweaks, but that's from a purely time perspective. The amount of time it would have taken us to actually copy the data was nothing compared to the hour of downtime that we actually had to do the head shelf. From that perspective, from a user’s perspective, everything paid off quite well. From the admin side, there was a little bit more work than just turning off a switch and flipping a switch.

Speed is obviously one of the most valuable features, with IOPS being the most important for certain applications: database applications and so on; performance with certain applications that has blown away the benchmarks set by the providers.

From an IT perspective, providing that as a platform for these specific databases has made us seem like gods, in short. There is a perception of IT in our organization that we're not capable or we can't provide the services that the other departments want. When they come to us and we give them space on AFF, they're blown away by the performance, as are the people that are telling them, "No, you guys can't provide it. Use these guys or go with a cloud provider." We're more than capable technically and now more than capable technologically.

I don't know if it's going to be possible in the short term to improve upon it because the drive technology is developing much faster than the processing technology, the CPU, that sort of thing. In the future, I'm sure they'll tackle that but right now, drive technology is accelerating.

We’ve had it for something like a year and a half.

We’ve purchased a second larger system and we've had no stability problems with it.

If one can afford the drives, then they're scalable. That's the caveat. Of course, there are some issues with scalability that come from the ability to crush your controller with so many drives behind it. If you have too many disk shelves, you can overwhelm a controller, one of the lower-end controllers. That’s a potential problem. It's not a problem we actually have, but it's something we have to be careful with because we have a mid-range AFF, and now we have an enterprise AFF as well. Now that we have the enterprise AFF, this isn’t an issue.

I have used technical support through a provider, C-Store. They were great.

I was consulted during the decision process to invest in the All-Flash FAS. It was an addition to FlexPod. We were told we needed to have flash storage to support an application when the truth was, we didn't actually need flash storage but there was an edict to do that. We went with the AFF in order to support the demands of a customer. We're happy enough with it to buy another.

In small ways. For the AFF, I was involved in the initial setup but not directly doing a whole lot of it. I consulted, and we set up the aggregates and all that based on specifications. It was straightforward and, again, we had good providers; good help makes things easy.

EMC was a possibility. I'm sure there were others. For me, it wasn't even a competition. I would have just said, "We're going with NetApp. We can talk about who's going to provide it but NetApp's the way to go." We were so heavily invested in NetApp already; also, most of our storage team had experience with NetApp and bringing in another storage vendor... learning curve and all that; we're already understaffed and over-utilized.

When selecting a vendor to work with, they have to be able to both support and anticipate our needs, communicate efficiently and clearly. Sometimes that means making changes in the way they do business in order to facilitate our needs because we have very little movement in the way we do business. We're a public school, a lot of stakeholders. We are beholden to explain ourselves to a lot of people. Those kinds of criteria are very important. Whatever we're buying has to be worth the money because we're not going to get it again very soon.

Make sure that an AFF is necessary before you buy one, because a FAS full of SSDs is very expensive and might not be necessary to meet your needs. You get plenty of IOPS out of a SAS and they are comparatively inexpensive so that you can increase your spindle count to make up for the IOPS of SSD; when you do that, you gain capacity too.

Don't let yourself be bullied by a vendor saying, "This software solution requires this level of hardware to back it up," because NetApp has already proven that's not the case.