Marquette University didn’t have an effective disaster recovery process in place because of the high cost and IT time required to support one. But after upgrading to Windows Server 2012 R2, which includes enhanced replication management features
and works with Windows Azure Hyper-V Recovery Manager, Marquette was able to create a cost-effective solution that can recover critical workloads in minutes rather than hours. The new solution offers reassurance to university management, staff, and students.
Marquette is a Jesuit university in Milwaukee, Wisconsin, with approximately 8,100 undergraduate and 3,700 graduate and professional students from every US state and 68 countries. It has more than 1,100 faculty members.
The Marquette data center primarily runs Microsoft software, and the IT staff likes to stay current with the latest offerings from Microsoft so that the department functions as cost-effectively as possible. Marquette was an early adopter of the Windows Server
2012 operating system, which it used to reduce storage costs by 10 times, trim server costs by 50 percent, increase IT staff efficiency, and reduce maintenance-related downtime.
The university had workload protection built into its data center by using Windows Server 2008 R2 cluster failover technologies, but the IT team wanted to provide better disaster recovery services to protect workloads in the event that an entire data center
was disabled. The university had a second data center just a few blocks away from its primary data center, where it ran staging and development workloads. If something catastrophic were to happen to the primary data center, the IT team would get workloads
running in the second data center as quickly as possible. But the effort was a “scramble” rather than an orderly process.
||By using Windows Server 2012 R2, we finally have a solid, highly automated disaster recovery solution in place for the cost of our Microsoft Campus Agreement.
| Anthony Ciske
Windows Administrator, Marquette University
“We needed a stronger disaster recovery process,” says Anthony Ciske, Windows Administrator for Marquette University. “We’ve had some near-disasters in the past that were a real pain to recover from. We needed a real disaster recovery solution that was compatible
with our budget—and our staffing bandwidth.” The team had explored commercial disaster recovery solutions in the past, but they were too expensive for serious consideration.
When Microsoft announced the Windows Server 2012 R2 operating system, Marquette took advantage of its Microsoft Campus Agreement to deploy an early version of the new operating system and participate in the Microsoft Rapid Deployment Program. The IT
team was specifically interested in testing the disaster recovery and multisite replication enhancements of Windows Server 2012 R2. Hyper-V Replica has been enhanced to support a third replication site, to replicate hundreds of virtual machines at a time and
to support replication frequencies down to 30 seconds.
Hyper-V Replica in Windows Server 2012 R2 also works with Windows Azure Hyper-V Recovery Manager, the Virtual Machine Manager component of Microsoft System Center 2012, and Windows Azure to provide simplified, cloud-based management of the replication process.
The control panel of Hyper-V Recovery Manager is delivered as a cloud service that runs in Windows Azure so that, in the event of a disaster, the disaster recovery orchestration metadata is securely accessible from any location.
Marquette worked with Microsoft Services Consulting to create two server clusters (using HP ProLiant blade servers), each of which represented a site. The IT team connected these clusters to Windows Azure Hyper-V Recovery Manager to test the multisite replication
process. “Microsoft Services helped us overcome a few challenges in getting linked up with Windows Azure and getting our first replicas set up,” says Aaron Ott, System Administrator at Marquette University. “It was nice to be able to talk directly to the product
group to resolve issues. The failover was quick—just a few minutes of downtime and then the workload was up and running in the other data center.”
The IT team can create a recovery plan based on Hyper-V Recovery Manager to orchestrate the disaster recovery process and replace most manual steps using scripts. “I really like the ability to nest a series of actions in the recovery process,” Ciske says.
“For example, we can orchestrate that each tier of an application—web, application, and database—restart in a certain sequence so that the web tier doesn’t start before the database tier. This helps us create an orderly recovery process with all our assets
kept in a consistent state.”
Marquette is currently strengthening its formal disaster recovery plan by prioritizing workloads for recovery, and ordering the hardware necessary to move its Hyper-V disaster recovery solution into production. Beyond disaster recovery, Marquette plans to
deploy every application that supports Windows Server 2012 R2 on the new operating system.
At the same time, Marquette upgraded to Microsoft System Center 2012 R2, which it uses to monitor the replication process and manage its entire IT infrastructure. The IT team uses the Virtual Machine Manager component for routine virtual machine provisioning,
the Operations Manager component to monitor server and application health, and the Service Manager component to manage IT service requests. It uses the Configuration Manager component to deploy software to desktop and server computers and is just starting
to use the Orchestrator component to automate IT workflows.
By upgrading to the latest Microsoft data center software, Marquette University gained a comprehensive, cost-effective disaster recovery solution that it can use to protect university workloads in the event of a catastrophic disruption to its primary
data center. The solution provides greater assurance that workloads and data are protected.
“By using Windows Server 2012 R2, we finally have a solid, highly automated disaster recovery solution in place for the cost of our Microsoft Campus Agreement,” Ciske says. “Hopefully we’ll never have to use it, but we have peace of mind in knowing that
if something terrible happens, we won’t have to engage in a mad scramble to recover our workloads.”
Workloads that previously took days to recover can now be restored in minutes. Plus, Marquette gained a disaster recovery solution that uses its current infrastructure and is familiar to staff, rather than a third-party solution that costs US$60,000 and
requires a major learning curve. “That’s huge for us, because we are a small staff with more work than we have people to do,” Ciske says. “Every efficiency that Microsoft software provides is put to good use around here.”
For more information about other Microsoft customer successes, please visit: