Windows Server 2008 R2
Customer Solution Case Study
/ Hospital Saves $260,000 and Achieves Critical High Availability with Virtualization
Overview
Country or Region:United States
Industry:Healthcare
Customer Profile
Greenville Hospital System (GHS) is the largest healthcare system in South Carolina, with five medical campuses, 100 clinical buildings, and more than 10,000 employees.
Business Situation
The hospital faced server sprawl, underutilized servers, and aging server hardware. It wanted to not only consolidate servers, but also find a highly available, cost-effective server solution.
Solution
GHS decided to virtualize its server environment by using Windows Server 2008 R2 with Hyper-V. It uses the Microsoft System Center suite of products to manage its virtual and physical assets.
Benefits
  • Saved at least US$260,000
  • Delivered high-availability systems
  • Streamlined IT maintenance
/ “Using the $5,000 average cost each for 52 servers, we have already saved $260,000 by using Windows Server 2008 R2 with Hyper-V and virtualizing, instead of using a one-to-one ratio for physical machines.”
Patrick Meindl, Senior Systems Engineer, Greenville Hospital System
Greenville Hospital System University Medical Center (GHS) provides healthcare services across South Carolina. Technology plays a key role at GHS, providing the critical systems that providers need to care for patients. In order to create highlyavailable, cost-effective systems that not only keep the hospital viable, but also help ensure regulatory compliance, the organization implemented a virtualization strategy built on Window Server 2008 R2 with Hyper-V and Microsoft System Center products. By January 2012, GHS eliminated 25 physical servers, avoided acquiring 52 servers, and saved at least US$260,000. The organization has a highly available infrastructure that is not only cost-effective, but easier to maintain. Impressed with the results, GHS will continue its virtualization efforts and plans to virtualize almost 50 percent of its infrastructure, representing 300 virtual machines.

Situation

Founded in 1912, Greenville Hospital System University Medical Center (GHS) has grown from a single hospital with 84 beds to the largest healthcare system in the state of South Carolina. GHS now has five medical campuses and more than 100 other clinical buildings, including outpatient centers, radiology centers, and physician practices—all spread across the entire upstate region of South Carolina. The hospital system has more than 10,000 employees and thousands of contractors that provide care to patients, promote healthy living, and work to advance healthcare for future generations.

Like any hospital, GHS has to comply with government mandates, including regulations such as the American Recovery and Reinvestment Act of 2009 (ARRA) and the Health Insurance Portability and Accounting Act of 1996 (HIPAA). Complying with these, and future, regulations introduces challenges for the hospital’s IT department. Also, it is a delicate balancing act to provide both providers and patients with easy, yet secure, access to electronic health records.

Given these needs, technology plays a critical role at GHS, providing clinical staff withaccess to the information that they need to provide quality care and patients with self-service capabilities, such as the ability to schedule appointments online. The IT department at GHS uses a centralized infrastructure to support its geographically dispersed healthcare system, including more than 600 servers in two centrally-located data centers. It is this server infrastructure that must be highly available, because hospital systems need to run 24 hours a day. “Clinical outcomes are closely tied to our technology infrastructure,” explains Patrick Meindl, Senior Systems Engineer at Greenville Hospital System. “Patients need our services around the clock, so we have to make sure that our IT environment is also ready around the clock.”

In addition to providing a reliable infrastructure, the IT department must also run the infrastructure in the most cost-effective manner possible. “Senior leadership has done a tremendous job managing costs in this economic climate, and it’s something that the IT department watches carefully,” says Meindl. “To remain viable and provide quality care, we have to keep our costs down.” However, three issues threatened the hospital’s ability to provide a highly available, but cost-effective, infrastructure: old hardware for a portion of its server fleet, server sprawl, and underutilized servers.

First, a growing number of the servers at GHS were nearing end-of-life, running on older, less-efficient hardware, and needed to be replaced. GHS initially identified at least 12 servers that needed to be replaced within a short timeframe, at an average cost of US$5,000 each, but as high as $8,000 each, depending on the server configuration. It knew that an increasing number of servers would require replacement in the future, too.

Second, GHS experienced server sprawl as a result of the high number of test systems that it creates. “Every time a department needed a test system for a new software application or clinical system, we would purchase a new server for testing purposes,” explains Meindl. “The problem with that is that you have to not only spend the capital to purchase server hardware, but then you also have to spend money on running the server, cooling the hardware, and maintaining these systems.”

Finally, the hospital also had several systems that it considers low-risk decision-support systems—systems on smaller servers with dual-core processors and 2 gigabytes (GB) of RAM that only five or six employees use in a day. “These smaller servers only have 5 percent utilization at the heaviest point during any day,” explains Meindl. “There’s no reason to have these systems running on a dedicated piece of physical hardware—it’s a waste of resources.”

Solution

In November 2008, Greenville Hospital System decided to explore server virtualization as a solution for its underutilized sprawl of servers, hoping to cost-effectively consolidate many of its servers and replace older hardware without incurring capital expenses. The organization initially chose the Windows Server 2008 operating system with Hyper-V technology for its virtualization solution, along with Microsoft System Center technologies.

GHS started with a proof-of-concept project and implemented four blade servers with 146 GB of RAM each, running on Windows Server 2008 with Hyper-V. On each server, it ran three virtual machines, for a total of 12 virtual machines for the initial proof of concept. It used those virtual machines as a starting point to virtualize any new test systems that were requested from various departments across the hospital. The IT department used a third-party recovery product to make image backups of each virtual system without operating system requirements—ensuring that business users would have a point-in-time copy of the image. “We wanted to make sure that users were comfortable with virtualizing critical systems during this proof-of-concept phase of the project,” says Meindl. During the proof of concept, GHS implemented Microsoft System Center Virtual Machine Manager 2008 to manage its virtual machines. With System Center Virtual Machine Manager in place, the IT department used the physical-to-virtual migration capabilities available in Hyper-V to move servers to a virtual environment.

After the successful proof-of-concept project, GHS decided to evaluate all requests for new servers, including those for test and production environments, and use virtualization technology for all systems that did not require dedicated, standalone physical hardware.

At that point, GHS decided to run heavier workloads in a Hyper-V test environment and targeted a resource-intensive medical application from Siemens for virtualization. The application requires two database servers—one for the primary application and one for back-end data analysis tasks—as well as two front-end application servers and a training server. For that system, GHS used a HP DL580 server with 85 GB of RAM running Windows Server 2008 with Hyper-V. “Putting this application in a virtual environment was a major milestone for GHS,” explain Meindl. “It is a complicated, resource-intensive system that handles a heavy volume of user traffic throughout the day. It performed extremely well running on Hyper-V, despite the load that it has to handle.”

Next, GHS moved from a test phase to a production phase. It used another HP DL580 server with four quad-core processors and 100 GB of RAM to run 50 virtual machines, again using the physical-to-virtual migration capabilities in Hyper-V and System Center Virtual Machine Manager 2008. The virtual machines on that physical hardware ran low-risk production systems, including various application servers and several Microsoft SQL Server 2008 databases used by a small number of employees.

In late 2009, GHS upgraded to Windows Server 2008 R2 Datacenter and set up two, four-node clusters for its high-risk production systems. It migrated servers that were previously running the original release candidate of Windows Server 2008 to promote a high-availability architecture. It also moved several critical hospital applications from General Electric (GE) and some critical database applications to the Windows Server 2008 R2 clusters. GHS took advantage of Cluster Shared Volumes, a feature of Failover Clustering introduced in Windows Server 2008 R2, which the organization uses to share storage resources among the 77 virtual machines running between both clusters. One cluster has two virtual hosts and 43 virtual machines, the other has two hosts and 30 virtual machines. By the end of 2009, GHS had eliminated 25 physical servers by adopting its virtualization strategy.

The hospital system also upgraded to System Center Virtual Machine Manager 2008 R2 to manage its virtual assets. It implemented Virtual Machine Manager Self-Service Portal 2.0, which it uses to give application support staff the ability to perform routine tasks through a web-based portal. In addition, GHS uses Microsoft System Center Configuration Manager 2007 R2 to manage its physical IT assets, including servers and client computers, and to keep an accurate inventory of its virtual machines. The organization uses Microsoft System Center 2012 Operations Manager to monitor the health and performance of its IT environment, and plans to upgrade to Microsoft System Center 2012 by the end of 2012.

GHS also plans to continue its virtualization efforts into the future. For its server environment, it plans to virtualize almost 50 percent of its resources. “We expect that instead of using more than 600 physical servers, we will have closer to 350 physical servers and 300 virtual machines,” says Meindl. In addition, the organization looks forward to the next release of Windows Server, codenamed “Windows Server 8.” Moving beyond its server virtualization efforts, GHS is also evaluating a virtual desktop infrastructure (VDI) solution built on Windows Server 2008 R2 with Hyper-V, in addition to Microsoft Application Virtualization (App-V) and Citrix XenApp. “This is just the beginning of virtualization for GHS,” says Meindl.

Benefits

As a result of implementing a virtualization strategy built on Windows Server 2008 R2 Datacenter, in addition to Microsoft System Center technologies, Greenville Hospital System has a healthy, highly available server infrastructure that helps it meetgovernment mandates and patient and provider needs. By virtualizing its environment, GHS has eliminated aging servers, avoided the need to purchase additional servers, and saved at least $260,000. GHS expects the savings to increase significantly as it continues its virtualization efforts.

Reduced Server Acquisition and Licensing Costs

GHS consolidated server roles, thereby eliminating 25 physical servers initially. It also expects to eliminate a total of 300 physical servers in the future. “Those are servers that we no longer have to license, power, and cool, which is a tremendous savings,” says Meindl. In addition, out of the 77 virtual machines that it deployed initially, 52 of those were new systems, for which GHS would have previously obtained new hardware. “Server prices range from $3,000 to $8,000, depending on the application requirements,” explains Meindl. “On average, though, our servers cost approximately $5,000. Using the $5,000 average cost each for 52 servers, we have already saved $260,000 by using Windows Server 2008 R2 with Hyper-V and virtualizing, instead of using a one-to-one ratio for physical machines. That savings will continue to rise as we move from an infrastructure that is 20 percent virtualized to one that is closer to 50 percent virtualized.”

In addition, by using Windows Server 2008 R2 Datacenter, GHS takes advantage of flexible licensing for compounded savings, and itcan run an unlimited number of virtual machines for each host without incurring additional licensing costs. “If we had to license each virtual machine, that would create significant overhead,” says Meindl. “Unlike other virtualization technologies, byusing Windows Server 2008 R2 Datacenter, we can avoid expensive licensing costs. We can get a lot of saturation on one physical box and really make the most of our investment.”

Delivered Highly Available Systems to Meet Strict Compliance Regulations

By using Windows Server 2008 R2 with Hyper-V, GHS now has high availability in a single data center. In the event of a system going offline, virtual machines can fail over to a different node and the hospital will not lose any of its critical systems that providers rely on to provide patient care. “High availability is extremely important from a compliance standpoint—and from a care standpoint,” explains Meindl. “Our patients need our services 24 hours a day, sevendays a week and, by using Windows Server 2008 R2 and Hyper-V, we can better guarantee that our systems are always running.”

GHS reports an increase in server uptime across both of its server clusters. “We have a critical system that thousands of clinical staff rely on and if it goes down, the IT department gets complaints almost immediately,” explains Meindl. “After implementing Hyper-V and System Center Virtual Machine Manager, no one notices if a virtual machine fails over. For example, we had a networking issue with a single host and had to fail over the virtual machines on that one host to a different host immediately. End users were not impacted, which was particularly important in this case because the application is sensitive enough that one request time out is extreme enough to shut down the application.”

Streamlined IT Maintenance and Health Monitoring

By using Windows Server 2008 R2 with Hyper-V, in addition to System Center products, the IT department at GHS has streamlined server maintenance and improved the overall health of its server infrastructure. For instance, the organization configures servers to automatically go into a maintenance mode one Sunday each month on an alternating schedule. Once maintenance mode starts, the virtual machines from one host start failing over to another host. The host in maintenance mode then automatically installs any critical updates and restarts automatically. The next morning, an IT administrator moves the virtual machines back to the updated host without any downtime.

“If we had to manually migrate individual virtual machines back and forth to update a host, it would be a time-consuming and tedious process,” says Meindl. “By using Hyper-V and Virtual Machine Manager, we can automate much of the process and ensure a healthy, load-balanced infrastructure that is always available.”


Windows Server 2008 R2

Windows Server 2008 R2 is a multipurpose operating system designed to increase the reliability and flexibility of your server and private cloud infrastructure, helping you to save time and reduce costs. It provides you with powerful tools to react to business needs faster than ever before with greater control and confidence. For more information, visit: