A CIO’s Perspective on HPE SimpliVity: There’s Nothing More Valuable Than Time
Data is the lifeblood of today’s businesses. But what happens if there’s a major catastrophe—a fire or flood—and you need to recover and restore your entire environment?
No problem, right? Every CIO has a plan in place to address these kinds of situations. So, the question then becomes, not can you restore your backups, but how long will it take? How many hours or days of business will you lose? How much money will it cost you in missed opportunity?
These were the questions that led me to look for an alternative to my traditional SAN setup. My company, Tanner LLC, is the largest locally owned accounting firm in Salt Lake City. Our clients depend on us to handle—and protect—their most sensitive financial data, so I knew that while we were already getting the benefits from high availability and redundancy from the current VMWare and SAN solution, I knew we could do better. I needed to be able to perform offsite backups and restore the environment without hours of delay.
The Limitations of SAN
With our old system, I didn’t have a good way to get backups off the SAN. If I needed to restore, I would have to operate at a file-level and folder-level, rebuilding the machines by recovering the most recent, successful folder-level backups. For a large server or multiple servers, this is complex and time consuming task. It would’ve likely cost us several days to get everything back up and running, provided we could get new hardware in place.
I was also struggling with our storage system. All of our employees run multiple accounting apps, which use a large amount of data, and all of this data is stored on the network. Because we didn’t have deduplication and compression, our SAN was running out of space. We were at about 90% capacity on our HP LeftHand. There was no deduplication and compression capability built into it; it was just raw storage. So, basically, any time someone would write a file, if we had two identical files, it would take up double the space.
The whole process was extremely stressful, as well. If I heard someone needed to store a bunch of new data on the network, I would have to think, "Well, what's this going to do to the overall storage capacity?" We needed a smarter solution.
Evaluating Our Options
I’d had enough of the worry and inefficiencies of our old solution, so I began to evaluate our options based on three major factors: price, reliability, and architecture. I asked a few peers at other organizations about their setup, but after hearing from a former colleague about hyperconvergence and the benefits he was realizing from this new technology, I had to learn more.
It was 2014, and hyperconvergence was rather new. A value-added reseller (VAR) introduced me to HPE SimpliVity (this was actually before the HPE acquisition). At the time, they were an up-and-comer, and didn’t have an extensive track record. But their presentation blew me away.
My hesitation about their age as a company was quelled by the fact that they were building their system on well-known hardware. At the time, they were building on Dell servers, so that was important. One of their competitors was using rebranded white box servers, so the HPE SimpliVity servers seemed more enterprise tested. Then they talked about the R&D hours that went into developing their hardware accelerator card for storage and about the extensive engineering, and that’s when they gained my trust. These guys new their stuff.
When it came to features, the built-in deduplication, built-in backup and restore, and the ability to do off-site backups with ease struck a chord with me. It was exactly what I needed. I liked the completeness of the solution. Then we discussed price.
If I had pieced a solution together from multiple vendors, the cost of separate storage, separate servers, and then the software on top of it would’ve been close to or exceeded the price of the integrated HPE SimpliVity offering—only without the seamless communication of one unified solution. Plus, many of the other competitors we evaluated were priced out of our range. They weren’t even hyperconvergence, more like convergence, and they were priced more for large enterprise.
Another vendor we considered was Nutanix, but it, too, was beyond our reach and their feature set didn’t even come close to HPE SimpliVity. Here, I’ll give you an example: HPE SimpliVity uses the hardware accelerator card for storage operations, so the deduplication and compression happen on the fly. They happen in-line. And it also applies to any network traffic, as well, so anything between VMs is also deduplicated. Whereas Nutanix performed those data operations post-process, so you would have to go back through and use CPU and RAM and all that to do deduplication after the fact. Whereas the HPE SimpliVity server didn't even touch the CPU. It was all done in the card before it even wrote it to disk or flash. That is a superb feature in the HPE SimpliVity server appliance architecture.
Also, I didn't fully realize it at the time, but where that really has the advantage, where the rubber hits the road, is in the backup and restore operations and all the data operations you do on VMs. Copying a VM, cloning a VM, backing up a VM, restoring an entire VM—it all happens in seconds.
I don't think any other vendor can do these operations as fast. Needless to say, I was sold on HPE SimpliVity.
One concern with any major technology purchase is implementation. So often it takes days and countless headaches to just get something up and running. With HPE SimpliVity, what we experienced was the exact opposite. In fact, it was pretty much done in just one afternoon—remotely.
The local VAR who sold us the system sent one of their senior engineers to our site to observe, to basically just watch. He had never seen an HPE SimpliVity implementation and was curious to how the whole thing worked.
As the implementation went along—a remote HPE SimpliVity engineer handling the entire process—the VAR engineer, who was an expert on VMware, kept asking, "Wait. Don't you have to do this or that?" The HPE SimpliVity guy said, "Nope. It's all automated. You don't have to worry about a lot of the minutiae in how you provision storage. It does it automatically."
It took a bit for the VAR engineer to get his head wrapped around that. It's just one giant storage blob, and HPE SimpliVity manages it for you.
"Well, don't you have to go in and carve out storage for each VM?" he asked. "No, you don't have to,” said the HPE SimpliVity engineer. “That's the beauty of it. It's just really easy." It really was that easy.
Putting It to the Test
A few months after implementation, I had a chance to see exactly how powerful our new solution could be. A few of our internal users came into my office and said they were having application issues with a program that uses a network share hosted on 300 GB VM guest server running on HPE Simplivity.
I went to my console and started looking. I could see that the files were messed up—nothing was right on it—and so the best way to fix it was to just restore it. In the past, what that entailed was going to our file and folder-level backup system, which was a Barracuda backup server, and then just choosing the folders and bringing them back down. But due to how extensive the issues were on this app, I figured that the easiest way, the quickest way, rather than trying to restore all the files, was just to restore the entire 300 GB server.
So, I decided to just use the built-in restore functionality, because I had been getting really frequent backups already defined in HPE SimpliVity. I just hadn't restored any of them yet, so I wasn't sure what to expect. I right-clicked on the VM, and clicked "restore." The restore steps took two or three clicks in the VMware interface. Then I turned away from my computer to tell one of the users that it would take a while, and he needed to go tell the rest of the department that it might take an hour or longer. By the time I turned back around to my screen, the 300 GB server VM showed it as restored. I couldn’t believe it.
I thought, "There's no way it finished that quick. Restoring an entire server in under 20 seconds?” I wondered if I missed some steps or something. However, after powering on and booting the server, we found that everything was back to normal. The entire outage lasted less than 10 minutes, and of that time, the server restoration operation was under 1 minute, with only a few minutes of lost work due to the frequent backup schedule. We all came away very impressed with the real-world trial by fire that could have been very costly on another system.
I really love and count on the built-in data protection and recovery in HPE Simplivity. The backup, restore, and off-site backup capability is top-notch. I have complete confidence that our critical data is protected while HPE Simplivity is on the watch.
In addition to the backup and restore capabilities, I also have off-site backup. When I bought HPE SimpliVity, that was one of the big things that drew me in. I wanted to have a way that I could somehow send backups off-site, and since we have a single site, I knew that we required a colocation site.
What I ended up doing was renting a quarter-rack space in a data center that's about 300 miles away from our office. Then I set up a basic 20 Mbps VPN connection between our office and that data center to have that remote partial server rack simply be an extension of our network over the VPN.
When I bought HPE SimpliVity, I bought two HPE SimpliVity systems to go in our main office, so they have copies of each other. They're configured in a redundant VMware high-availability cluster. Then I bought a third system that was going to be the backup system, and I put that down at the remote data center. Whenever I do backups, I do them very frequently here locally, and then I also do them every couple of hours to send complete backups of every single VM I have down to the remote data center 300 miles away.
In systems I've worked with in the past, it usually took a long time to back something up over a VPN. But with HPE SimpliVity, depending on how much data is changed on a server, they are usually backed up in only a few minutes. The most I've seen it take is about 15 minutes, even for a server that's heavily used. So, every couple of hours, I'm getting the most up-to-date copies of our data stored off-site.
Now our entire operation is 100% virtualized on HPE SimpliVity. There's really not anything, production-wise, that runs outside of HPE SimpliVity. It handles everything, from file, print, our production applications, file storage, user home directories, all our accounting apps like the CCH Prosystem fx Suite—you name it, it's running on HPE SimpliVity. Our security products. Our entire operation, literally, is on two servers. Two HPE SimpliVity servers. And it runs it perfectly.
The best thing? It's still got room to grow. I can keep adding more apps to it, and I don't worry about it suffering at this point. Today, it's hovering around the 50–60% utilization range. I've still got some room to grow, and that, to me, has been the best part of this entire experience.
HPE SimpliVity is, by and large, set it and forget it. I can go a whole week or longer without ever even looking at any statistics on it. I'm not worried about storage filling up, and I'm not worried about backups failing. It's just kind of out of sight, out of mind. It is nice to not have to worry about the day-to-day configuration, monitoring, statistics, and shuffling data around. Instead, I have my time back and can focus on bigger projects, and for a busy CIO, there is nothing more valuable than time.