We’ll explore how adopting storage innovation protects the state’s largest hospital from data disruption and adds operational simplicity to complex data lifecycle management.
To describe how more than 150 terabytes of data remain safe and sound, we’re joined by Jeff Bergholz, Manager of Technical Systems at The Nebraska Medical Center in Omaha. The discussion is moderated by BriefingsDirect’s Dana Gardner, Principal Analyst at Interarbor Solutions.
Here are some excerpts:
Gardner: Tell us about the major drivers that led you to seek a new backup strategy as a way to keep your data sound and available no matter what.
Bergholz: At Nebraska Medicine, we consist of three hospitals with multiple data centers. We try to keep an active-active data center going. Epic is our electronic medical record (EMR) system, and with that, we have a challenge of making sure that we protect patient data as well as keeping it highly available and redundant.
We were on HPE storage for that, and with it, were really only able to do a clone-type process between data centers and keep retention of that data, but it was a very traditional approach.
A couple of years ago, we did a beta program with HPE on the P6200 platform, a tertiary replica of our patient data. With that, this past year, we augmented our data protection suite. We went from license-based to capacity-based and we introduced some new D2D dedupe devices into that, and StoreOnce as well. What that affords us is to easily replicate that data over to another StoreOnce appliance with minimal disruption.
Part of our goal is to keep backup available for potential recovery solutions. With all the cyber threats that are going on in today’s world, we’ve recently increased our retention cycle from 7 weeks to 52 weeks. We saw and heard from the analysts that the average vulnerability sits in your system for 205 to 210 days. So, we had to come up with a plan for what would it take to provide recovery in case something were to happen.
We came up with a long-term solution and we’re enacting it now. Combining HPE 3PAR storage with the StoreOnce, we’re able to more easily move data throughout our system. What’s important there is that our backup windows have greatly been improved. What used to take us 24 hours now takes us 12 hours, and we’re able to guarantee that we have multiple copies of the EMR in multiple locations.
We demonstrate it, because we’re tested at least quarterly by Epic as to whether we can restore back to where we were before. Not only are we backing it up, we’re also testing and ensuring that we’re able to reproduce that data.
More intelligent approach
Gardner: So it sounds like a much more intelligent approach to backup and recovery with the dedupe, a lower cost in storage, and the ability to do more with that data now that it’s parsed in such a way that it’s available for the right reason at the right time.
Bergholz: Resource wise, we always have to do more with less. With our main EMR, we’re looking at potentially 150 terabytes of data in a dedupe that shrinks down greatly, and our overall storage footprint for all other systems were approaching 4 petabytes of storage within that.
We’ve seen some 30:1 decompression ratios within that, which really has allowed my staff and other engineers to be more efficient and frees up some of their time to do other things, as opposed to having to manage the normal backup and retention of that.
We’re always challenged to do more and more. We grow 20-30 percent annually, and by having appropriate resources, we’re not going to get 20 to 30 percent more resources every year. So, we have to work smarter with less and leverage the technologies that we have.
Gardner: Many organizations these days are using hybrid media across their storage requirements. The old adage was that for backup and recovery, use the cheaper, slower media. Do you have a different approach to that and have you gone in a different direction?
Bergholz: We do, and backup is as important to us as our data that exists out there. Time and time again, we’ve had to demonstrate the ability to restore in different scenarios, the accepted time of being able to restore and provide service back. They’re not going to wait for that. When clinicians or caregivers are taking care of patients, they want that data as quickly as possible. While it may not be the EMR, it maybe some ancillary documents that they need to be able to get in order to provide better care.
We’re able, upon request, to enact and restore in 5 to 10 minutes. In many cases, once we receive a ticket or a notification, we have full data restoration within 15 minutes.
Bergholz: We’re pretty much all dedupe-type devices. It’s not necessarily SSD, but it’s good spinning disk, and we have the technology in place to replicate that data and have it highly available on spinning disk, versus having to go to tape to do the restoration. We deal with bunches of restorations on a daily basis. It’s something we’re accustomed to and our customers require quick restoration.
In a consolidated strategic approach, we put the technology behind it. We didn’t do the cheapest, but we did the best sort of thing to do, and having an active-active data center and backing up across both data centers enables us to do it. So, we did spend money on the backup portion because it’s important to our organization.
Gardner: You mentioned capacity-based pricing. For those of our listeners and readers who might not be familiar with that, what is that and why was that a benefit to you?
Bit of a struggle
Bergholz: It was a little bit of a struggle for us. We were always traditionally client-based or application-based in the backup. If we needed Microsoft Exchange email boxes we had to have an Exchange plug-in. If we had Oracle, we had to have an Oracle plug-in, a SQL plug-in.
While that was great, it enabled us to do a lot, it we were always having to get another plug-in thing to do it. When we saw that with our dedupe compression ratios we were getting, going to a capacity-based license allowed us to strategically and tactically plan for any increase that we were doing within our environment. So now, we can buy in chunklets and keep ahead of the game, making sure that we’re effective there.
We’re in throes of enacting archive-type solution through a product called QStar, which I believe HPE is OEM-ing, and we’re looking at that as a long-term archive-type process. That’s going to a linear tape file system, utilizing the management tools that that product brings us to afford the long-term archive of patient information.
Our biggest challenge is that we never delete anything. It’s always hard with any application. Because of the age of the patient, many cases are required to be kept for 21 years; some, 7 years; some, 9 years. And we’re a teaching hospital and research is done on some of that data. So we delete almost nothing.
In the case of our radiology system, we’re approaching 250 terabytes right now. Trying to backup and restore, that amount of data with traditional tools is very ineffective, but we need to keep it forever.
By going to a tertiary-type copy, which this technology brings us, we have our source array, our replicated array, plus now, a tertiary array to take that, too, which is our LTFS solution.
Gardner: And with your backup and recovery infrastructure in place and a sense of confidence that comes with that, has that translated back into how you do the larger data lifecycle management equation? That is to say, are there some benefits of knowledge of quality assurance in backup that then allows people to do things they may not have done or not worried about, and therefore have a better business transformation outcome for your patients and your clinicians?
Bergholz: From a leadership perspective, there’s nothing real sexy about backup. It doesn’t get oohs and ahs out of people, but when you need data to be restored, you get the oohs and ahs and the thank-yous and the praise for doing that. Being able to demonstrate solutions time and time again buys confidence through leadership throughout the organization and it makes those people sleep safer at night.
Recently, we passed HIMSS Level 7. One of the remarks from that group was that a) we hadn’t had any production sort of outage, and b) when they asked a physician on the floor, what do you do when things go down, and what do you do when you lose something? He said the awesome part here is that we haven’t gone down and, when we lose something, we’re able to restore that in a very timely manner. That was noted on our award.
Gardner: Of course, many healthcare organizations have been using thin clients and keeping everything at the server level for a lot of reasons, a edge to core integration benefit. Would you feel more enabled to go into mobile and virtualization knowing that everything that’s kept on the data-center side is secure and backed up, not worrying about the fact that you don’t have any data on the incline? Is that factored into any of your architectural decisions about how to do client decision-making?
Bergholz: We have been in the throes of desktop virtualization. We do a lot of Citrix XenApp presentations of applications that keeps the data in a data center and a lot of our desktop devices connect to that environment.
The next natural progression for us is desktop virtualization (VDI), ensuring that we’re keeping that data safe in the data center, ensuring that we’re backing it up, protecting the patient information on that, and it’s an interesting thought and philosophy. We try to sell it as an ROI-type initiative to start with. By the time you start putting all pieces to the puzzle, the ROI really doesn’t pan out. At least we’ve seen in two different iterations.
Although it can be somewhat cheaper, it’s not significant enough to make a huge launch in that route. But the main play there, and the main support we have organizationally, is from a data-security perspective. Also, it’s the the ease of managing the virtual desktop environment. It frees up our desktop engineers from being feet on the ground, so to speak, to being application engineers and being able to layer in the applications to be provisioned through the virtual desktop environment.
And one important thing in the healthcare industry is that when you have a workstation that has an issue and requires replacement or re-imaging, that’s an invasive step. If it’s in a patient room or in a clinical-care area, you actually have to go in, disrupt that flow, put a different system in, re-image, make sure you get everything you need. It can be anywhere from an hour to a three-hour process.
We do have a smattering of thin devices out there. When there are issues, it’s merely just replaying or redoing a gold image to it. The great part about thin devices versus thick devices is that in lot of cases, they’re operating in a sterile environment. With traditional desktops, the fans are sucking air through infection control and all that; there’s noise; perhaps they’re blowing dust within a room, if it’s not entirely clean. SSD devices are a perfect-play there. It’s really a drop-off, unplug, and re-plug sort of technology.
We’re excited about that for what it will bring to the overall experience. Our guiding principle is that you have the same experience no matter where you’re working. Getting there from Step A to Step Z is a journey. So, you do that a little bit a time and you learn as you go along, but we’re going to get there and we’ll see the benefit of that.
Gardner: And ensuring the recovery and voracity of that data is a huge part of being able to make those other improvements.
Bergholz: Absolutely. What we’ve seen from time to time is that users, while they’re fairly knowledgeable, save their documents where they save them to. Policy is to make sure you put them within the data center. That may or may not always be adhered to. By going to a desktop virtualization, they won’t have any other choice.
A thin client takes that a step further and ensures that nothing gets saved back to a device, where that device could potentially disappear and cause a situation.
We do encrypt all of our stuff. Any device that’s out there is covered by encryption, but still there’s information on there. It’s well-protected, but this just takes away that potential.
You may also be interested in:
- Loyalty management innovator Aimia’s transformation journey to modernized IT
- Infrastructure as destiny — How Purdue builds an IT support fabric for big data-enabled IoT
- Cybersecurity crosses the chasm: How IT now looks to the cloud for best security
- Big data and cloud combo spark momentous genomic medicine advances at HudsonAlpha
- How software-defined storage translates into just-in-time data center scaling and hybrid IT benefits
- Securing data provides Canadian online bank rapid path to new credit card business
- How Allegiant Air solved its PCI problem and got a whole lot better security culture, too
- How European GDPR compliance enables enterprises to both gain data privacy and improve their bottom lines
- DevOps and security, a match made in heaven
- Alation centralizes data knowledge by employing machine learning and crowdsourcing
- Expert panel explores the new reality for cloud security and trusted mobile apps delivery
- Catbird CTO on why new security models are essential for highly virtualized data centers