Cisco Blogs


Cisco Blog > Data Center and Cloud

Lack sufficient I/O infrastructure..of time to provision servers?Think UCS

May 21, 2010
at 12:00 pm PST

The IT department had already begun adopting server virtualization, using VMware ESX software on rack-optimized servers. But the new clinical data warehousing applications would require 16 additional VMware ESX servers, and the data center lacked sufficient I/O infrastructure and cabling. In addition, provisioning this quantity of physical servers would be difficult in the few weeks available”
Sounds familiar ? As I was talking with customers at EMCWorld in Boston, but also at SAPPHIRE in Orlando over the two weeks, listening to  customers from Chicago, New York, Frankfurt, London , Paris , or visiting UCS customers such as Pacific Coast Building Products Inc. , this concern seems to be all over the place.
The challenge of Moses Cone Health System, a large healthcare provider based in North Carolina is not unique, but nevertheless it is critical for the success of the IT organization’s mission.

“We needed a cost-effective computing system that would enable us to expand our use of Electronic Medical Records quickly over the next year, minimize network infrastructure build-out, and reduce time to rack and configure servers,” explained also   Michael Heil, manager of technology infrastructure, Moses Cone Health System. 

Amongst the numerous benefits provided by the acquisition of the Cisco UCS , Michael Heil was keen to highlight  :

 

  • Significant CAPEX reduction compared to the cost of traditional blade servers
  • Rapid and easy deployment  - Heil estimates that his team saved 1 to 2 weeks for all 16 servers.
  • Much better management “ Cisco UCS managers reflects changes far more quickly than most Java appllications”
  • Faster recovery
  • Simplified Disaster recovery
  • More flexibility

Although the Cisco UCS is optimized for virtualization, it can also support physical servers in the same chassis as VMware. Heil’s team took advantage of this flexibility when it received the unexpected news that the database servers for a new document management application had to be installed on physical servers. By hosting Microsoft Windows Server on the Cisco UCS, the Health System can take advantage of hardware abstraction to provide mobility for the physical server. “This simplifies our business continuity and disaster recovery plans,” says Heil.

As another example of the flexibility of the Cisco UCS, the IT team initially configured server blades with 96 GB of memory and then later realized that 48 GB would be sufficient. To remove memory, the usual process is to power down the server, remove the chips, and then reboot, a process that can take 15 minutes per blade, which is difficult to schedule in a 24-hour healthcare environment. The Cisco UCS shortened the process: a Health System engineer removed the memory from unused blades and then moved the service profiles to these blades. “Using service profiles in this way reduced downtime from 10 or 15 minutes to 4 minutes, and avoided the risk of a 1-hour outage if a memory chip was not seated correctly,” says Heil. “Service profiles also gave us confidence that the application would operate correctly as soon as it was moved.”

Heil’s conclusion ? “Cisco UCS has given our organization more flexibility and greater agility, giving us a solid foundation to meet changing healthcare industry demands. It’s an exciting time to be in healthcare, and tools like the Cisco UCS are helping us move forward.”    
” A solid foundation to meet changing industry demands ” That’s definitely   the best compliment you can expect from an IT customer!

Please take the time to read the one page case studie that Michael Heil put together, detailing more the operational benefits baked with interesting data and numbers, and contact me if you want to share  your experience as well.

 

Comments Are Closed