NetApp AFF - All Flash Reviews

Reduced latencies, and the cluster data ONTAP, less down time, able to do upgrades, things like that, without much disruption.

I would say the consistency with the ONTAP versions and the speed and performance from the flash.

A shorter list of bug fixes would make it a 10 out of 10 for me. It looks like they're doing monthly releases now, so there are a lot more upgrades. It feels like a little too much, but we get to choose whether or not we need to pick that version or if we're going to wait. It's good not to have to wait four months for a patch.

With some of the larger clusters being able to do a patch upgrade is helping. They still take three, four hours by the time you get the night started, finish things up, do the upgrade. The upgrades are very minimal. They've got the waiting period in between them, kills about 15 minutes of time. It'd be nice if that was streamlined a bit. I'm sure the engineers have that pause in there for a reason.

Normally good.

I think we've got an eight-node cluster right now, so it's meeting our needs.

It's been easy to tag nodes and scale out.

It's always been a good experience. I've never had any issues getting the right level of support.

Pretty straightforward.

I would say the primary use case for AFF is a combination of database and virtual servers. We have both block storage and file storage.

Our impression of NetApp as a vendor of high performance SAN storage, both before and after we purchased AFF, was top-notch. We are definitely more likely to consider NetApp for mission critical storage systems in the future based on our experience with AFF, due to its reliability, ease of administration, cost.

For us, reliability, cost, and just a good relationship are the most important criteria when selecting a vendor.

It's reliable, fast, low latency, and we haven't had any issues with it. It's been quality.

With the AFF, we can run VMs with databases now. That was one of the big features with the AFF, we needed the speed for databases. By moving them over, we can put VMDKs housing databases on there and use them on the VMware infrastructure now.

Learn about the benefits of NVMe, NVME-oF and SCM. Read New Frontiers in Solid-State Storage.

The AFF we have, we use the in-line compression. The in-line dedupe, and the compaction saves us a lot of space because most of our AFFs house VMware VMDK files. We got a lot of compression, a lot of efficiency out of the dedupe because a lot of the VMware are similar with the OS, VMDKs, etc. It makes it really compact. You can put a lot of stuff in a little space.

That's a hard question to answer off the top of my head. I'd have to go through and evaluate everything. Right now, it fits our needs. I'd have to evaluate what else I'd like to see, I guess.

While not for AFF specifically, for clusters in general, it would be nice to be able to have volumes everywhere. For example, now you have volumes tied to a node tied to an HA pair. It would be much better if you had it more like the way they do Metro Clusters, where they have a switch, and the storage is all attached to a switch. Then, they have a volume owned by something and have it should be able to move around to anywhere based on ownership of a volume, as opposed to between HA pairs. That would be a good improvement in their infrastructure.

The NetApp AFF itself, the FAS's, they're stable. They're in a cluster mode, they're HA, so we fail them over, we have upgraded fail back. We've never had an outage due to NetApp in the 12 years that I've been there.

Learn about the benefits of NVMe, NVME-oF and SCM. Read New Frontiers in Solid-State Storage.

Scalability, it's like anything else. The ability now to take out and add shelves, pull out shelves from the middle of an array if you want, to upgrade them, to pull heads out, and put new heads in as a non-forklift upgrade. All that functionality and scalability is one of the things that makes NetApp really good for our environment.

We use tech support for everything. Since it's a cluster, something that's not specific AFF, it's just nodes in the cluster. But we use support all the time.

Tech support is like everything else. It's hit or miss. It depends on who you get and what the subject matter is. We had a Support Account Manager (SAM) at one point too and, when we had the SAM, it was a lot easier to work with their support through the SAM. We've dropped the SAM stuff.

Sometimes it's difficult to escalate correctly and get the right people involved. It's not been as bad as it was before we had the Support Account Manager (SAM) though. Our SE helps a lot as well. It's pretty good support. We just had a support call yesterday with him and the guy we got was knowledgeable about what our problem was, so it worked out pretty well.

We've been a NetApp customer for 10 to 12 years now. We use their non-flash stuff a lot. We use hybrid flash, and after that, hybrid arrays. All Flash was the next logical move. Our next move is going to be the object storage, as well to spin off some of that data, the snapshots, on to object storage, because they've got flex groups.

I was involved and it was seamless. We had a two-node star cluster with AFAs on them. NetApp did the install. A few years ago, we used to do our installs ourselves, as a company. Then we started using NetApp installation services to do them. They did the install. They inserted it seamlessly into our cluster. It came up, we had the arrays, and we could create aggregates on it, pretty much right after they got them installed.

We're using NetApp now as our hybrid storage. We have VMs on there. They wanted to put databases on the VMs. We said, "Well, we don't have the speed to put your databases on there. If you want to stay on the NFS structure with NetApp, the next logical solution is just to put you on All Flash, so we just throw some of those in the cluster and do a motion of your volumes over."

For All Flash, we have a SAN infrastructure and a NAS infrastructure. We use the EMC for the SAN infrastructure, for the block. NetApp is the only NAS we have. There's not much else we can look at besides Isilon. Isilon just isn't fast enough. It's slower than what we had them on at the beginning. NetApp was really the only logical choice for that particular environment if we wanted to use NAS.

The primary use case for our All Flash FAS (AFF) system is pretty much VMware and its servers. It's just for file storage right now, for NFS, for the VMware stuff. We're investigating using it for other things. It's also used as a Zerto, a web application depository for some of the Zerto replication for the VMware stuff.

We use it for our mission critical stuff right now, as our VM infrastructure.

The most important criteria, when selecting a vendor to work with is functionality. I look at the functionality of the systems, what they provide us, what the features are, and where they're going, and what we need. Then, after that, I'll look at support. Of course, my company wants to look at market share and similar thing to it, but I look at the those things last. I look at the functionality first.

I give it a nine out of 10 because nothing's perfect. It works really well for what we want to do with it. It may not work well for other people. But in my experience, nine is where I would put it. It's functional, it's expandable, no forklift upgrades, and no disruptive upgrades, even for the OS or for the hardware itself. The flexibility of moving things around. All of its features, including its SnapMirror functionalities, make it really good for our environment.

All the features and their flexibility is where I would give it the bigger rating. What would make it a 10 out 10 is better support.

Regarding advice, it's the same advice you give to everybody. Evaluate what your criteria are, then look at NetApp. If you're looking for NAS, even for block, NetApp to me is mid-to-high level block. If you're looking for certain things in block, something else might be better, as opposed to FAS. You can look at NetApp for their other products. Look at NetApp for their file system for; FAS, look at their block stuff. Look at their stuff because all their stuff is available for use, it's just that the FAS itself is not suitable for everything, but they have other stuff that is.

The primary use case for our Flash FAS is general storage for our hypervisor, software as a service provider. We primarily use it for storing our applications, web servers, file servers, and whatever other applications we have. We mainly utilize the AFF platform for the high performance and the deduplication capabilities. The management is a lot simpler on an AFF.

• Deduplication
• Compressions
• Simpler management
• The performance is great.

I'm not sure there are any additional features which I want to see, except for maybe more compatibility within the hardware universe and more compatibility for cables and other hardware. Some better integration with the E-Series to give us more options to scale. The other issue though is a completely different product called HCI, so this might not even be an AFF request.

It would be helpful if the compatibility matrix was a bit better. That's what we run into a lot. Our datacenters have a need for more flexible cabling and NetApp has very strict guidelines on what kind of cables you use. That's the only reason why I wouldn't rate it a 10 out of 10, but everything else is great.

Stability is very good.

Regarding scalability, on a scale of one to 10, I'd say about a five.

Tech support is great. I'd rate it as a nine out of 10.

The initial reason for going to NetApp was that our original solution, which was Dell Storage, just wasn't cutting it. We did our own in-house testing, performance-wise, resilience-wise, etc. The Dell Storage just wasn't cutting it. Dell's other solution at that time was Compellent, and NetApp was just better. The initial reason we didn't go with NetApp was because of cost, but they were able to meet us in the middle and we just went from there.

Not straightforward, there is a learning curve when it comes to AFFs, but once you understand the setup it's pretty easy.

Our initial perception of NetApp was it's extremely expensive and a little too inflexible. However, once we did get into the NetApp ecosystem, we realized that the cost effectiveness was greater than we originally thought.

The cost effectiveness is due to deduplication compression, the number of managed hours that we need to maintain the system, and the flexibility of NetApp which is geared toward keeping their systems more resilient.

I would check to see that you're okay with centralized storage because that's what NetApp's bread and butter is. If you want a centralized storage platform that is bulletproof, NetApp is great.

We use AFF for both block storage and NAS storage. We are more likely to consider NetApp for mission critical storage systems based on our experience with AFF.

Two functions are valuable for us:

• Snapshot: We are able to replicate/clone production environments to test the SW version up (e.g. the Oracle Financials upgrades).
• MetroCluster: Our disaster recovery is based on two active-active sites. The MetroCluster feature allowed us to continue our operations (without business interruption) when we stopped one of the sites.

Learn about the benefits of NVMe, NVME-oF and SCM. Read New Frontiers in Solid-State Storage.

This solution is based on the scale-out concept. Some workload balancing activities across the nodes are not transparent (requires server downtime).

When moving volumes between controllers, you should always use the optimized path.

FCP is doing this automatically, but NFS unfortunately not.

So when moving NFS volumes between controllers, you will not move the load to the other controller.

To do this, you need to remount the volume to the correct LIF.

We did encounter stability issues but nothing that interrupted the solution; more background type of problems.

Learn about the benefits of NVMe, NVME-oF and SCM. Read New Frontiers in Solid-State Storage.

We did encounter scalability issues, the solution is not delivering the requested performance (I/O response time for the requested IOPS).

The technical support level is between poor to medium in our geography.

Previously, we were using the older generation of the NetApp MetroCluster (6240 unit). We switched as this unit was not performant enough, i.e., it had a high maintenance cost for the performance delivered.

Negotiate everything, i.e., including the price for the future capacity upgrades as part of the deal.

We looked at EMC, HPE, and Fujitsu.

You need to understand the limitations of the scale-out architecture.

Clustered Data ONTAP

Low latency

My company uses mostly NetApp products, so I have existing knowledge of using their products.

We're still testing it, but I expect it to provide us a lot more speed because of low latency and clustered Data ONTAP.

They still have to reduce in price when compared to their competitors. Also, the current version has some problems with global deduplication.

It's not yet in full production, but I have been testing the product for few month now with VMware.

I've had no issues deploying it in our test environment.

We do not expect any issues, and have had none so far.

It has problems with deduplication when done globally.

Their customer service is one of the best, and I am a demanding customer.

• Tier 1 – room for improvement as they hold onto tickets for too long
• Tier 2– much more serious

From my experience with NetApp products, initial setup is going to be nice and easy. We are very techy, so it was easy for us.

The majority of work was based on existing knowledge, but we also got help from the vendor.

We considered Kaminario and XtremIO. We chose NetApp in order to utilize current resources.

It has better adaptation than pure flash solutions such as XtremIO. It’s important to learn the weak spots of the suppliers in the market, and I can say that I have great expectations for the migration of the flash array to disc via cluster.

The customer improved its time to market.

Learn about the benefits of NVMe, NVME-oF and SCM. Read New Frontiers in Solid-State Storage.

I have found three main features to be valuable:

Ease of use: Business continuity solutions are not typically so easy to manage from a storage admin prospective

Storage Efficiency: Inline compression, inline reduplication, and other inline features allow space-saving without losing performance

Performance: All flash disks allow extreme performance at low latency

There should be more functionality regarding tiering of the oldest data.

The solution is very stable.

Learn about the benefits of NVMe, NVME-oF and SCM. Read New Frontiers in Solid-State Storage.

The solution can scale-in and scale-out.

I would rate the level of technical support 10/10.

The customer previously had the NetApp solution based on hybrid disks.
They don't have a business continuity solution.

It took one week for the startup to pepare the storage for use. We have migrated about 150TB of data in six months (VMware, Oracle, SAP, filesystem, etc.).

Licensing is very simple: all flash solutions include the entire license. Regarding pricing, storage efficiency can lower the cost per TB.

The customer evaluated EMC and HPE.

Involve a competent and certified partner.

Right now, we've seen a few different systems that we're running on the all-flash system, where we've seen performance increases with application functionality. We have databases running on there. The database query is running faster. The application is running faster in general. It has saved us by not having to tax the system to get the data access going quicker, less network usage. People using the applications are able to perform their tasks more quickly.

Learn about the benefits of NVMe, NVME-oF and SCM. Read New Frontiers in Solid-State Storage.

I'd say the biggest one for us, other than just being SSDs, was the compression; inline compression, inline dedupe. Previously, we used dedupe but compression in dedupe has helped a lot, just to be able to maximize our storage, not having to buy more disk and items such as that. That is the biggest one we've seen so far.

It’s difficult to say because there are already a lot of features that have been released that we didn't have previously, especially going from 7-mode to cDOT. ONTAP 9 sounds really interesting with better dedup and compression; the disk partitioning features that they are going to be doing with that. I'm eager to see what ONTAP 9 has. Right now, I believe were on 8.3, so we’re definitely going to be interested in upgrading that when it comes out.

I’m not sure if I see anything that's really lacking because there are so many features that we still have not taken advantage of that we could probably use going forward; no specific ones that I can see right now.

We've only had it about two or three months. We haven't had any issues since we've had it up. It's been in production and has been rock solid so far. I don't have a long-term say on that yet, but it's been really good so far.

Learn about the benefits of NVMe, NVME-oF and SCM. Read New Frontiers in Solid-State Storage.

We're not a huge shop. Our previous NetApps have always been a two-node setup. Right now, I don't really necessarily see us scaling out any more. We were pretty much a 7-mode shop previously; now, we're a cDOT for these 8080 AFFs. With cDOT, it's very nice how you can scale it out and add more nodes to it. I don't necessarily see us taking advantage of that anytime soon. It's nice to have the option there.

We've had 30-40 controllers for about five or six years now and we've previously had the NetApp 2000 series. We have kind of been a NetApp shop. We've had different vendors like Pure Storage previously come in just to talk about stuff. I think the main reason we went to All Flash was the price point.

When we were looking, we were doing a big project in which we were re-hauling a lot of our core infrastructure. We wanted to refresh the hardware on the NetApps. At the time, we were looking at doing a hybrid of spinning disk and SSDs; maybe doing flash pools and that kind of stuff. Then, working with our vendor and working with NetApp, we were going to need more space anyway so the cost of the new system plus additional shelves for the space was pretty much the same price at which they could give us an All Flash system. With the 4-to-1 compression and the similar features All Flash has to offer, it was kind of a no-brainer to move to that; a lot of performance increase as well, being on All Flash.

A lot of our workloads aren't really disk-intensive, so we don't really need all flash, so at the time it wasn't needed, but the price point that NetApp was able to bring it in at was a deciding factor. Also, at the time that we reviewed Pure Storage, a lot of our systems were using multiple protocols on the same controller; we were using fiber channel, NFS, CIFS. The Pure Storage systems, at least when we reviewed them at the time, they didn’t really support all of those protocols on the same controller. We would have to buy multiple systems to be able to cover all our protocols. That made them more expensive. That was definitely a disadvantage for them.

I was in charge of the original setup. I worked with our vendor to help actually do the install and configuration. It went really well. Coming from a 7-mode background to a cDOT was definitely a lot different with the lists and similar items to configure. It was very straightforward. We pretty much got it on the network within something like 30 minutes; got our VMware environment pointed to it and within a couple of hours, we already had data on it in the first half of the day.

I don't have a lot of experience with other vendors. We've reviewed Pure Storage, and even though we didn't officially have Nimble in, we've talked to Nimble at a lot of booths in some of the trade shows. They are pretty much the same as Pure Storage when it comes to some of their features, restrictions and similar items. EMC, I don't have any experience to speak for.

I've been using NetApp for a long time now, so I really like NetApp, especially with the new ONTAP features, with clustering going forward. Give a good look at NetApp. They have treated us well and their product has been really rock solid for us.

There's always a little room for improvement.

We have had no issues, but we never went to production.

The technical support is excellent. Anytime we've had any kind of questions, our rep can help us or we'll call into NetApp auto-support. We have not had any problems. Tech support is knowledgeable and their response times are good.

We compared this tool against EMC’s XtremIO head-to-head, and the NetApp blew it out of the water. There was no competition. We were already a NetApp shop, so they were our preferred tool anyway. It has more features and links to my OS, innovative CIFS, and deduplication. We had the knowledge of the system already. It wasn't reinventing the workforce.

The installation was pretty easy. It was my third setup, so it was nothing really new. There's only one minor switch that turns it into an AFF.

We use the system to do stuff that isn't quite out yet. We love to do some oddball things. We're one of the first to use NetApp shift to compete and migrate away from VMware. We didn’t run into any issues with it, and it beat the competition.

When looking for a vendor, it's usually value first, which is not the right way to do it. That's what it comes down to. The value and then next is feature set.