Dell SC Series Reviews

What we are using from Dell EMC is their server line and their storage line. We're pretty well embedded with Dell EMC.

We no longer have an issue, especially in the support arena, of wasted hours: getting support, waiting for support to arrive onsite. It's been a huge time saver for us.

The most valuable feature is the customer support. Everyone has a product that goes wrong. It's how well they handle it when it does, and Dell EMC has been tremendous at that.

In terms of managing the solution, they've gone from "not bad to manage" to "downright easy."

We have integrated the system with Exchange and vSphere, among others.

They need to make it a little cheaper.

I won't entertain any other vendor at this point.

In terms of scalability, so far, so good. We run a whole lot of different product lines. The ability to scale, at least within the product series, has been wonderful.

Technical support has been top-notch.

The only thing that made the setup difficult were the layers of project management. I know they're going through some changes as Dell and EMC merge, but they really need to take a strong look at how project managers are deployed and what their level of knowledge is. More often than not, I'm having to run those calls.

We've been able to take them and run them completely to end-of-life. For us, storage side, that's about five years. We consider that the best turnover rate for us, while for server hardware it's three years. With support contracts, they've all made it to the end of those time frames.

Licensing, especially on the storage line, could use some simplification. It's not terrible, but, for example, with the Isilon series, they've gone to completely a la carte. A la carte is very difficult to traverse, when it comes to what you need. It would be more beneficial, at least from my point of view as a customer, if they did it more like car companies do, where there are package lines. That would be my one complaint.

I always recommend them.

Regarding the purchasing process, we were working with a VAR and they helped loop us together with the sales engineers and usually, within a few hours of discussion, we were able to drill down on everything. We had a product solution figured out, and almost next-day.

The most important criteria when selecting a product like this are the reliability of the equipment itself and the support.

I rate the solution at nine out of ten. They've really got the customer service, especially in the support arenas, nailed. They make information-gathering on their solutions, especially during pre-sale, easy to do.

We use it for converged infrastructure.

It helps us streamline deployment of VMs.

Customer support could be improved. It takes five or six reps to finally get someone who can help us.

I would also like to see a better ability to capture driver firmware, compatibility, and upgrade metrics to be compliant.

The stability is fine until it comes to patching, and then we have issues. Whenever we have issues related to driver and firmware, it's a pain.

The scalability is good.

We switched because of cost.

The initial setup was straightforward.

We've used everything from HPE to UCS. We've used a lot of different stacks.

I rate the solution at seven out of ten. To be a ten it would need better support and better compatibility metrics across the firmware driver stack.

We use it as the OS drives for our VMs. Our is a hybrid solution.

It's good storage product for our C and D drives for the virtualization space.

It has good dedupe and compression. Also, the built-in data migration capabilities are pretty good, as is the federation. When we started migrating the workloads from different storage platforms, like NetApp and XtremeIO, it helped us in moving in that direction.

It would be great if they could increase the compression and dedupe ratio.

We haven't had any issues with stability so far.

It's scalable.

Technical support is good.

We had XtremeIO for the past three or four years and, prior to that, we had NetApp. I think the SC Series works better. We are pretty happy with it. In terms of performance with mixed workloads, the SC Series is pretty good. We don't see a lot of latency as we saw with NetApp. But I would say XtremeIO and SC are similar in that regard.

Most storage platforms are the same, but when it comes to the performance and dedupe, as I said, those were the main criteria, what we were after when we talked to Dell EMC. The relationship and trust are also very important.

The setup was straightforward. We used Dell EMC Professional Services, and they came in. A company like ours is pretty big, so we use the vendors to come in and install things because we don't want to mess them up. They gave us the development documents, etc., so we're really happy.

The ROI is good.

It is affordable. We need more discount, but it's good. Pricing is more important than improving on dedupe or compression.

We looked at Nimble and NetApp. We went with Dell EMC. We have a good relationship since most of our stuff is on VMware. When Michael Dell bought VMware we said, "Yeah, we should increase our usage," and, of course, we have been using EMC for more than a decade.

Research is important. Understand the product. Just like any other technology you use, research is the most important part.

We use it for VDI, mainly. In terms of performance, there were some difficulties to begin with, with a lot of different upgrades. It took a lot of time because we've got several of them. With all the upgrades done, it has run pretty smoothly.

Right now, we've just got one particular system on it, where we're just trying to test the waters to figure out if it's good because we use a combination of Dell EMC and Cisco equipment. So far, the Dell EMC seems to be doing pretty well. There are some applications that we've run where it appears that the Dell EMC would be a better solution.

We like the integration, the nodes themselves, the way that everything is integrated together. It seems like it's less hardware to keep up with.

Speed and performance seem to be good on the hardware side.

One option I would like to see is, when you're up on the view-screen, to be able to incorporate getting to what HPE calls the iLO, the Integrated Lights-Out. To be able to get that instead of having to go back and trying to find IP addresses and re-institute those would be good. It would be good to be able to that put on the initial splash screen.

As I said, when we first deployed it, there were a lot of upgrades that had to be done. We would have trouble with this or that, networking would fall, stalls, delays. We would have to run through another upgrade. We've got about 50 or 60 of these pieces of equipment, and to do an upgrade would take 30 to 40 hours.

We've had to coordinate that back and forth, as this is a live system. Since we got through that process - and it took weeks - it seems to have been a lot more stable. It's almost to the point now that we can say that we're pleased with it. To begin with, we were ready to throw it out the window.

We like the scalability. It seems like we can increase it and add to it when the need comes, to meet our growth.

In my position, I don't use the technical support. We do have other employees who, I'm sure, have used it. There are several times that we've called them. So far, we have not had complaints about the tech support. They've been able to resolve the issues that we've had.

We have found it to be affordable. Comparing the cost with other hardware, it seems to be right in line with the compute and the storage that we get with it. The cost is a saw-off compared to other options because there's not a lot of difference between them. You just have to figure out what would be the best.

I would consider Dell EMC to be one of the top options, by all means. I've liked HPE, but it seems that they are integrating with that, as far as VMware is concerned. The two that we're currently using are Cisco and Dell. They're the main two that we're bouncing back and forth with.

At times the SC Series has been slow. Most of the time, we have found the problem is on the appliance side, not the hardware side.

I rate the solution at eight out of ten. To get it to a ten it would need fewer upgrades and getting things right to begin with. That's really one of the core issues that we've had with it: so many upgrades. Once every two to three weeks we've had to upgrade firmware or something else. Although now, that has slowed down.

It's the storage backbone for our virtual environment. So far the performance has been very good.

• Replication
• Uptime
• I/O

I would like to see higher compression, dedupe, faster I/O, and bigger drives.

I have no complaints, so far, about the stability.

It's very easy to scale.

I have not needed to use technical support yet.

We had old technology. We wanted to move to stuff that's more VMware-aware.

Our most important criterion when selecting a vendor is price.

We have found the solution to be somewhat affordable. The scalability seems a bit more expensive than even buying a brand new one so far. That has been kind of a drawback.

Do your research, check out all the vendors, and get your hands on it as much as possible.

It back-ends our vSphere virtualization farms.

Overall, it has really helped us to virtualize a lot of workloads where server or application owners were very hesitant to move away from their physical boxes because they were used to having local disks and the performance that came with that. With the SC Series SAN, the performance that we've gotten out of the boxes alleviated anyone's concerns. We do not get complaints about the performance of our virtual infrastructure.

Also, with auto-tiering, it's easier to understand than most arrays, knowing that all of your writes go to the tier that you specify, with easy-to-create storage profiles.

• Reliable
• Good performance
• Economical
  

We also utilize replication. And the data progression feature gives us the ability to economically utilize those solid state discs by having a three-tier solution. That was key for the product.

We make use of VMware's VMotion features quite a bit. You don't even know that the SAN is there. It just moves the workload. It does it quickly. We've been happy.

I would like to see an integrated key manager in the controllers. Currently, it's an external product. It would be nice to have the option of having a built-in key for self-encrypting drive features.

We haven't had any issues affecting the stability.

It was easy to expand. In year two of our owning the SC8000 SANs, we expanded the flash tiers and it was a piece of cake. Very easy.

I have called the SC Series support and they've been great.

I worked with a VAR on the initial setup of the product. It was very straightforward, very easy to bring it online.

We were able to afford two tiers of flash storage at a price where the competition was giving us one tier or just a handful of discs.

We looked at several vendors. I work in public education so we have to consider alternate sources for every major purchase. Dell EMC gave us the most bang for the buck and we are comfortable working with them as a company.

When we look to work with a vendor, the important criteria are follow-through, ease of communicating with the vendor, and the quality of technical support.

You have to do your preparation and research but that's no different than buying any SAN. Get to know your local Dell EMC pre-sales engineer quite a bit because they'll be able to help you properly size the unit.

In our mixed workloads, we generally see sub-10ms latency on the product. I don't really have any stats for high-end IOPS because we are not really doing high-performance computing. We have mixed workloads and I'm more concerned with latency than IOPS. But the performance has been great. We have been very happy with it.

It's easy to expand. Back in the day, if you wanted to expand, you would have to buy a block of disk from somebody or a new shelf. Here, if you have open spots, you buy some more disk and you expand.

What I really like, from the model line starting with the 3000 all the way up, is the flexibility. You can have spinning disk, you can have flash, you can have a combination. 

Another valuable feature is the performance of the auto-tiering. It will move hot data up to your fastest Tier 1 or move your slow data down. Data progression is what it's called. With the auto-tiering you can have multiple tiers, you can have your Tier 1 be either spinning or flash, all the way down to 7.2K. It will change the RAID on the fly so your writes come in at RAID 10. After they sit for a while, they get converted to RAID 5, then they'll cool off and move down the tiers. Your performance is kept going, while the cold data is moved to your slow, non-performance tiers.

With federation, you can have multiple systems across sites. You can treat them as one, and with a live migration, volumes don't go down. You can move them from site to site, doing maintenance, and keep your environment up.

They already integrate with Dell Storage Manager, so you can manage multiple, you can set up replication, you've got monitoring, vSphere, Hyper-V.

Going into the unit itself, to manage, takes a little more intuition. But when you integrate it into the Dell Storage Manager it gives a much more GUI, user-friendly area to manage, provision volumes, etc.

It's very stable.

It's very scalable, especially with federation. If you outgrow the number of spindles that a unit can support, depending on the model, you can go into federation.

Copilot support is next to none. I've dealt with a lot of support before. With Copilot, you're usually on the phone if you have a problem. Even as a customer, not a partner, if you have an issue or you're planning on upgrading code, you call Copilot, they do a system check, they give you the thumbs up. The support is very helfpul.

The 3000 Series are user-installable but, as a certified installer, I have found that customers, if they're not used to the interface, like to come in and have a tutorial to see the interface and how it's done. Anything in 5000 Series and above requires a certified installer. You want to make sure that you get your networking and zoning set up properly; and to get a walk-through on setting up the environment and for getting used to it.

It does not have a steep learning curve. Especially in the SMB, the front-end is not hard. You set up the Storage Manager, it gives you the dashboard and tells you when you're over-provisioned, and shows performance and the like.

It's very affordable. Anything from an SMB on up to the large enterprise - from a 3000 Series all the way up to a 9000 Series - with federation and you name it.

You should definitely have a partner come in. If you're going to do a multi-tier environment, look at your IOPS: What are your hot IOPS, what's your total storage need? You need to plan out those different tiers because that's where it saves you money. You don't have to go all-flash if you don't need it. You only need to be able to deliver your performance and most places have tons of cold data that they aren't aware of. So, having someone come in and do an assessment of your current storage environment and see what performance you really do need - what you're getting now and what your projected growth will be - is important, so that a system can be properly sized. That should be a pre-sales process. Your engineer should understand Compellent and be able to size it properly.

I just deployed a 3000 Series in a small VDI environment and, during a bunch of huge data copies, I was seeing 15,000 IOPS at less than about 7 or 8 milliseconds of latency and that was on 10K disk. I was really impressed with that.

The most important factor when looking at a vendor is, are they there to sell you a box or are they there to help you? Are they there for the long term?

From my dealings with it, it's right up there at ten out of ten. Obviously, there are more expensive systems out there but, for all the deployments that I've seen or done, its been a rock solid platform.

It's primary storage. It started off as just being spinning disk, but recently we've had a lot of traction with customers adding flash and even going to all-flash systems. Some of our customers have had it six, seven, eight years.

It performs quite well. We had a customer who had an Oracle all-flash system, and they needed to run VMs in their disaster recovery facility and, of course, they did not want to purchase the Oracle storage because it's extremely expensive. So, they added flash to their existing Compellent storage, they were using a DR for the other stuff. It has performed quite well for them.

One of the nice things about it is that there is no forklift upgrade required to upgrade the storage. That's why a lot of our customers like it.

The maintenance is usually pretty good. It's not like some of the others where they increase it in the fourth, fifth, or sixth years. That's another reason the customers like it.

We've been able to maintain a lot of customers over the last six or seven years. We actually started selling the SC Series - it was called the Compellent - before Dell purchased them.
We had picked it as a strategic storage for our company to sell. It has been good to us over the years. We continue to make a lot of sales.

The most valuable feature is the no-forklift upgrade. While the thing is running, I can change out the controllers one at a time and keep the customer up and running. I can add shelves and storage and SSD drives or spinning drives to the system, while it's running. I can bring all that in and rebalance the load across the new disks or, if we take disks away, rebalance the load across what's remaining, and it just works.

Also, in terms of auto-tieriing, Compellent writes all of its writes into Tier 1, unless of course you do something silly and pin a whole bunch of LUNS, which means you're telling your VM, your data stores, that they have to live in that top-tier storage. As long as you follow the best practices recommendations and let it do its own auto-tiering then it works very well. 

In most cases the customers with all-flash, most of their active data lives in flash. So, they're really using all of those IOPS and performance in that tier, and the other tier or tiers are just being used for cold data storage. It works very well, as long as customers follow the best practices there.

We can definitely see a need for it being a multi-controller system for customers who want to scale beyond the current capability. That's always a downside. A lot of the new systems are scaling vertically, they scale out, and the Compellent, of course, is controllers with shelves under it, so you don't scale out with it, unless you add another one. But if you do, they don't talk to each other, like some of the other solutions that we sell.

It's absolutely stable. There were some issues in the early days with the tiering, but they straightened that out. What was happening was, if the first tier filled up - and it wasn't necessarily because of pinning - but if the first tier filled up because you overran the "write" amount that you had in the first tier, it died, actually.

Dell EMC was really quick in fixing that. Now, if it gets to 90 percent, if you start to overrun that tier, it will start moving stuff automatically. It doesn't have to wait for the tiering, "seven down and three up." The way that it works is, if you're in the top tier here and something is not accessed in the last seven days, it will move down to the next tier and it takes three days to move it up. But with it automatically tiering it down, it can automatically go and take it out of there if it sees that it's going to fill up.

It scales very well. It's not a multi-controller system right now. They are limited to two controllers, like some of the others. It is an active-active controller, so both controllers are processing LUNS. One of them is called the leader and it is the transcriber - think of it as the "court reporter" guy. But both of them actually take reads and writes.

I do use technical support quite often. In fact, any time I do an upgrade I call Copilot and let them know that I'm going to be doing an upgrade, in case anything happens and they get an alert. They'll also do a system check for me so that I know everything's working at 100 percent before I go in there and start pulling cables or adding cables. It gives me peace of mind. A customer can't say, "Well, you broke it." It was that way before we started. And in some cases, support won't let me start the upgrade, they'll advise me not to. I work with them quite a bit. They are probably the best support that you'll find.

But recently, things have changed a little bit. I have gotten some wrong answers from Copilot. I have to double-check their work on certain things. I've had at least three times where I've gotten the wrong answer from someone, which is not good.

But they're always responsive, whether you call at three in the morning or three in the afternoon. It's the same group of people up in Minnesota. They're not in India, they're not in the UK, they're not in Turkey, they're not somewhere where they have a thick accent and you can't understand them. 

I think the reason for the changes, from what I'm seeing and what I've heard, is that they have now started pulling some of the other storage people, who may have done the EqualLogic stuff, the lower-end storage, into the Copilot fold, and are having them do the support. I'm sure it's just a popularity thing.

We were concerned too, when EMC came on board, that Dell would bury the SC line at some point. What we've been told is that they're going to continue it on. It's a much better product than even they imagined. It seems the product is going to stick around for a while, from what we've been told by Dell EMC.

We still sell some EqualLogic. I had worked with that before. I had also worked with the EMC Series for many years, back to 2000. I still do both EqualLogic and SC Series.

I'm a certified SC Storage Engineer and when our company sells the SC Series, I will go out and install it. I've been doing it for a long time so I can do a lot of the stuff with my eyes closed now.

It's not something that an end-user is going to pull out of the box and set up. You have to some knowledge of the SAS chaining, how it works with the Compellent, and how you load the software and how to set it up. It does have a pretty easy GUI on the setup. It's not command-line-based like some other stuff. The SAS chaining is the hardest part to understand.

We just had a customer that ripped out a whole cabinet of spinning disks and they're down now to one shelf of 24 flash disks and three shelves of spinning disk to replace all that. We're talking about a huge return on power, cooling, rack space, and cost of maintenance, because they do charge by the shelf for their maintenance. Now they've got four shelves as opposed to 18 or 19 that they had before. That maintenance cost goes down quite a bit because of that and that's all because of the SSD drives now, being able to add those in.

About four or five years ago, they finally brought on the what we call the hybrid and full SSD to existing customers. Initially, it started off that only new systems could be all-flash, and now we're going back and we're seeing a lot of our customers who add the flash in and revitalize their whole storage system, completely.

It's absolutely affordable. Again, not having to do a forklift upgrade, like some of the others have had to do in the past, it's very affordable for our customers and they continue to buy.

We have looked at different solutions over the years. We represent quite a few: EqualLogic, the SC Series, NetApp, and, now that HPE has Nimble, we sell Nimble. About the only ones we don't sell are Pure and Tintri.

We looked at a lot of others, back in the day, that are gone now. Going with the EqualLogic and the SC Series was a wise choice because they're still here, seven or eight years later and people are still buying them and people are still upgrading.

The SC Series is not the thing for your mom and pop shop. It is an enterprise-grade, mid-tier business type solution. It depends on how much space you need. Some situations can be fine with the EqualLogic series, although that's going to go away and, perhaps, the lower-end Unity's will replace them. You have to look at all those factors.

In terms of performance with mixed workloads, generally, the Compellent has been very good. If it's over the 20ms mark that's not good, but I have never seen any problems with that, in particular, when we've added flash.

In terms of migrations, I haven't used the built-in capabilities. I'll use VMware's VMotion to move things from one SAN to the other. A couple of years ago I had a Compellent upgrade where, at that time, they didn't support the upgrade I mentioned earlier, where you do one controller at a time on that thicker model, so I had to set them up side by side. They had some physical servers and I just used replication to replicate the LUNS over to the new SAN. I then shut down the physical servers for a few minutes and pointed everything to the new SAN and that worked great. I know they do have the data import, it's just something that we don't generally use because most of our customers are fully virtualized.

We sell the equipment and install it. I do both sides of the field, the engineering, both pre-sales and post-sales.

We still have a lot of customers that are in the 6 series. A lot of our customers have moved up to 7.1. If you have the SSD drives, then you can do things like dedupe, which you can't do on the older versions, and you can't do if you don't have the SSD drives. Evidently, the deduplication uses some part of the SSD drives for metadata or something like that, so that's a must.

There's some new stuff coming with 7.3, which just got released, where they're spreading the sparing across the whole array, rather than having a dedicated spare disk, and having it sit there and do nothing until one of them dies, and then it kicks in - and having to rebuild all of that. Now, they'll do the sparing across all the disks and they say that is not only going to add space but performance to the array. I've got a couple of customers that want to use that very soon.

I like the SC Series. I enjoy working with it. We've done a lot of sales lately, which is kind of surprising with all the new, fancy, all-flash arrays out there. Customers really don't need all-flash today, and that's really where SC fills in, with its auto-tiering, hybrid - mixed SSDs and spinning disks - customers just don't need all-flash systems. They don't have that kind of workload.