Case Study: Gateway Health

Gateway Health ^SM focuses on serving the needs of Pennsylvania’s most vulnerable citizens — those facing chronic diseases, the disabled, aging, and financially disenfranchised individuals. Established in 1992 as an alternative to Pennsylvania’s Department of Public Welfare’s Medical Assistance Program, Gateway Health provides its members with quality services for disease management, health and wellness programs, and preventive care. Today, Gateway Health is a top-ranked managed care organization providing service to more than 300,000 of Pennsylvania’s most vulnerable, at-risk citizens.

“Our business is very information-intensive,” explained David Guilinger, Director of Information Services at Gateway Health. “We have a higher number of claims per member and collect more data than traditional insurers. Because we are growing very fast, we also have the challenge of exponential growth in our business and the amount of data we must manage.”

Challenges and Goals − Life Before Nimble

Gateway Health had been relying on NetApp storage arrays for over six years, but the performance of the existing platform had started to degrade. “We implemented NetApp Flash Cache a few years back and saw some dramatic performance gains in our datacenter environment by leveraging flash,” Guilinger explained. “It worked well for read-specific tasks, but we were struggling with significant write-specific challenges. We tried some unsuccessful workarounds, like implementing a flash pool, but it didn’t solve the write issues.”

Gateway Health’s IT team also wanted to implement a flash pool for its rapidly growing Oracle-based database warehouse environment. “Not only did we have rapidly increasing capacity and processing requirements, we were running out of windows to process all of our data. We already have a plan in place to invest in Cisco UCS to improve compute performance. We wanted to see what could be done on the storage side as well — to improve write performance, not just the reading of data.”

Guilinger noted that there were two big drivers for moving to VDI. “The first driver was business continuity. In case of an epidemic, our people need to be able to work from home. The other was providing secure, remote access to all of our key business applications. Previously, our employees used encrypted company laptops and accessed our corporate environment using VPN. It worked some of the time, but there were several applications that didn’t work over those VPNs, even with broadband.”

Performance Issues in the Citrix VDI Environment

In addition to trying to improve the capabilities of the existing storage environment, Guilinger was also trying to address I/O challenges in its new Citrix virtual desktop environment. Gateway Health implemented Citrix XenApp and XenDesktop five years ago, with enough capacity to potentially support 500 concurrent users.

Even though Guilinger and his team followed best practices when implementing the Citrix environment, they were still not happy with the number of concurrent users they could support. “We have a lot of telecommuters,” Guilinger stated. “On any given day, approximately 200 of our employees are using virtual desktops. We were starting to get a lot of complaints about the sluggishness of the system. And it certainly wasn’t because we didn’t have enough virtual servers or memory. It was obvious that the bottleneck was I/O.”

Moving to Nimble Storage

Working closely with SynergIT, Gateway Health started looking for a new storage solution better suited to its VDI environment. “Several of our colleagues told us to check out Nimble Storage,” Guilinger said. “Our Nimble rep told a very compelling story — about where the founders and leadership came from, their approach of writing the CASL architecture from scratch, and how they examined the issues that were causing traditional storage vendors to struggle in their attempts to create a better solution. After a short POC on the Nimble arrays, we were convinced it was the right way to go.”

Gateway Health purchased two Nimble CS-Series arrays, one for its production environment and a second array for the company’s backup datacenter. “We decided to move our large Citrix environment to Nimble first, since we were receiving an increasing number of degradation complaints from users in our virtual desktop environment,” noted Guilinger. “After seeing how well the Nimble arrays worked in our VDI environment, we moved our MS SQL and Oracle environments on to the Nimble arrays as well.”

Cutting VDI Login Times by One Third

Nimble clearly delivered on its promise, according to Guilinger. “By deploying the Nimble arrays, we have eliminated all Citrix-specific latency from our VDI environment. The only latency that still exists is in the application — there is no added latency from the virtualization stack. As a result of moving to the Nimble arrays, we were able to cut initial login times to bring up a virtual app by one third.”

Gateway Health can now easily support 500 concurrent users without pushing the systems. “The end-user experience has really improved,” noted Guilinger. “Our users are pretty quiet when things are running well, but we hear from them immediately when they are not. We haven’t had any users complain about VDI performance since deploying Nimble arrays.”

35-50% Improvements in Oracle and SQL Performance

Gateway Health relies on an Oracle 11G environment with approximately 4 TB of data. “Since deploying the Nimble arrays, we are seeing an average of 50 percent performance improvements for all complex read / write processes in our Oracle environment,” Guilinger reported “Stored procedures that used to run for over six hours, now run for just 2.5 hours.”

“We know we have a few stored procedures that could be more efficiently written,” Guilinger admitted. “We were just getting ready to invest IT resources in recoding them, but there has been such a gain in performance with Nimble that we don’t have to take that project on right now. Instead of rewriting existing code, we can focus our time and energy on creating entirely new processes that will replace the legacy processes when we have time.” In addition to the improvements in the Oracle performance, Gateway Health saw 35 percent performance gains in its MS SQL environment for both read and writes.

Improving Data Protection and Disaster Recovery

Gateway Health’s main datacenter is housed in a co-location facility outside the city of Pittsburgh. The company’s backup center is located 25 miles away from the main center. The two geographically dispersed centers use different power systems and are connected using a fiber network. Gateway Health is using Nimble Storage for snapping and mirroring data across the two centers. This includes a ‘cold standby’ of the VDI environment that can be spun up when needed.

“We tried using NetApp SnapManager for Oracle, but we had a failure with that software at least once a month,” Guilinger reported. “We also plan to take advantage of some of the more intelligent cloning that comes with the Nimble arrays. We’d like the ability to keep more copies of snaps in our backup environment and production environment. That’s a Nimble feature we are looking forward to implementing very soon.”

Accurate and Timely Information from Nimble Storage InfoSight

Guilinger indicated that he uses Nimble Storage InfoSight three or four times per week. “InfoSight was easy for my team to adopt. It provides a lot of rich information about our arrays. We previously leveraged NetApp Operations Manager, but there were fallacies in the data it was presenting. When we would look at the amount of available storage, it was convoluted from what was really available. We certainly don’t have that problem with InfoSight. The information is easy to access, timely, and always accurate.”

Achieving Cost and Datacenter Footprint Savings

Nimble has also enabled Gateway Health to reduce its datacenter footprint. “We achieved a 50 percent reduction in datacenter space due to the small form factor of the Nimble arrays,” noted Guilinger. “Right before we switched to Nimble, I allocated an additional rack for storage expansion in both of our datacenters. But as I take out the older NetApp arrays, I can focus those extra racks to support our growth.”

“The cost to acquire the Nimble arrays was significantly less than buying more rack shelves from NetApp storage,” according to Guilinger. “Although the reason we went with Nimble Storage was to support VDI and improve application performance — a great side benefit for us was the dramatic cost savings compared to our previous storage platform. Better performance at a lower price is a rare combination for any IT solution!”

About Nimble Storage

Nimble Storage believes enterprises should not have to compromise on performance, capacity, ease of use, or price. Nimble has developed the first hybrid storage architecture engineered from the ground up to seamlessly integrate flash and high-capacity drives. Our customers enjoy as much as 10x faster application performance, enhanced backup and disaster recovery, and stress-free operations — all while lowering their TCO.

Nimble Storage solutions are available through a global network of world-class channel partners. Follow us on Twitter: @Nimblestorage.