Virtualization isn't just a buzzword, it's a technology that can potentially help to save thousands of dollars for CIOs and enterprise IT managers. To properly harness the economic benefits of virtualization, furniture retailer Slumberland teamed up with Cisco Systems and its Unified Computing System (UCS) platform.
The move to UCS-based virtualization has saved Slumberland's CIO approximately $368,000 in capital costs that might otherwise have been required to scale and support the business. With 118 stores and $390 million in annual sales, Slumberland is one of the top 25 furniture retailers in the U.S.
Slumberland CIO Jamie Page explained to InternetNews.com that he tracks a number of key metrics to determine how well IT is performing. Those metrics include system availability and total cost of ownership on a per-user basis, as well as ensuring that trouble tickets are answered within the bounds of their service-level agreements.
"On a bigger perspective, we're always watching the total cost of ownership, making sure we have the appropriate margins with our gross margin and return on investment," Page said. "I'm involved with marketing, merchandising and the logistics folks."
With all those tasks in mind, Slumberland also only has two system administrators to run its network. As a result, the company required a virtualization technology that was easy to acquire and learn, so it chose Microsoft's Hyper-V virtualization as the underlying virtual machine base over market leader VMware.
"We just found that the things that VMware does really well, beyond what Hyper-V does, are things like disaster recovery and fully automating that and memory oversubscription, and we decided we didn't really need those things," Seth Mitchell, IT manager at Slumberland, told InternetNews.com. "We needed only 80 percent of what VMware does, and Hyper-V covered everything quite well."
Slumberland decided to deploy Hyper-V on Cisco's UCS as part of its overall network transformation away from an older generation of Dell blade servers. The decision to move to the new hardware was part of an overall lifecycle-planning process.
"We had older hardware that we needed to replace to ensure reliability and performance and also a lot of new service demands that we had to address at the same time," Mitchell said.
When it came to choosing a hardware platform, Mitchell explained that Slumberland didn't do a bake-off against other hardware vendors. He said that memory is the key enabler for the company's applications, and in his view, other vendors did not have the same kind of memory capabilities that Cisco offered with the UCS.
The Slumberland UCS deployment is made up of two UCS compute fabrics that both have two UCS chassis. Within the chassis are a total of 10 UCS B200 M1 half-width blades and six UCS B250 M2 full-width blades.
Prior to UCS, Slumberland had 95 servers. Now, the four UCS chassis provide more capacity, take up significantly less space and are easier to manage with UCS, Mitchell said.
From a migration perspective, Slumberland leverages Microsoft System Center physical-to-virtual tools for the conversion of existing equipment. Provisioning the new servers was accomplished via Hyper-V and Windows deployment services.
In terms of best practices for deploying UCS and virtualization, Slumberland has one key suggestion.
"Don't be scared off by the idea that it's a new technology or that it's a lot of Cisco Kool-Aid," Mitchell said. "It's a lot of conventional technology that is applied in a novel way."
Deployments such as Slumberland's use the same fibre-channel interfaces as non-UCS systems, and will work with existing systems.
"It's really simple and it's a good way of abstracting servers," Mitchell said. "We do this kind of thing in IT every day with DNS, storage and other things and it works great. Why someone didn't do it with servers before -- I don't know."