Data Center Energy Efficiency Solutions

The traditional “one workload, one box” approach to server provisioning inevitably leads to over-provisioning and underutilization of server hardware assets. IT organizations typically provision at least one server for every application or service they deliver, yet most of these servers only operate at about 5-15% of their total load capacity. Meanwhile, the costs associated with power, cooling, network infrastructure, storage infrastructure, administrative overhead and real estate can skyrocket as the data center expands. The result is server sprawl: a situation in which servers take up more space and consume more resources than their workloads can justify. With escalating performance of server components, expanding IT footprint within the organization and the adoption of high-density form factors, many data centers have maxed out their power and cooling capacity.

VMware Customer Case Studies

To illustrate how TCO methodology can be applied to real scenarios, VMware spoke with three customers who shared details about their VMware Infrastructure implementation and how it affected their power and cooling costs. This sample of VMware customers represents a range of industries, company sizes and implementation scenarios.

Case #1: 1800Radiator Reduces Power and Cooling Consumption by 25%

Two years ago, Radiator Express Warehouse (1800Radiator), a fast growing automotive-parts distributor, began franchising. While initial growth was slow and the company had time to add infrastructure and build servers, in the last year, its business skyrocketed. The company has grown to more than 130 franchise locations at a rate of three new openings per week.

“Before we knew it,” says CTO Mike Carvalho, “our computer room was almost at maximum power capacity, our computer racks were full, our switches were all used, and our air conditioning was on continuously. Rather than spend thousands on new power systems, racks and air conditioning, we chose to leverage VMware’s product line to allow our company to keep pace with our growth.”

To maintain its high service levels, 1800Radiator turned to VMware and partner AccessFlow. Using virtualization technology, the company was able to take 31 physical servers out of production—the benefits from which extended beyond the combined savings of a nearly $6,000 PG&E rebate check and 25-percent reduction in power and cooling costs. “It was worth it to me just for the peace of mind in being able to support and grow my business. VMware High Availability (HA) and DRS, for example, provide automatic redundancy and resource management such that we can maintain maximum uptime—simply and affordably.”

Results:

• Avoided purchase of $100,000 worth of new hardware
• Removed 31 physical servers, eliminated the “one app per server” rule, improving flexibility and availability
• Reduced power and cooling consumption by 25 percent in the data center, saving $5,000 per year
• Primary and disaster recovery site both managed from single VirtualCenter console
• Reduced downtime due to hardware failure and maintenance
• Primary and secondary data centers both run below 50-percent capacity, allowing instant failover support of entire computing load
• Received a PG&E rebate of close to $6,000 for consolidation Project

Case #2: QUALCOMM Saves $1.87 Million in Space, Power and Cooling Infrastructure

QUALCOMM, a leader in the digital wireless communications field, is best known for pioneering and commercially developing Code Division Multiple Access (CDMA) and other digital wireless technologies. More than 470 million consumers worldwide rely on CDMA for reliable voice communications and data services.

“We are a VMware-first company. Our internal customers receive virtual machines as the rule. Our allocation model is based on resource needs, not server needs, and provisioning a physical server is the exception,” says Paul Poppleton, senior staff engineer. As resource needs change, we add capacity through just-in-time hardware provisioning and retirement of unused virtual machines. In three years, our admin count in the Windows server group has remained flat, while the OS-instance count has more than doubled. The only explanation is virtualization.”

Results:

• Increased server utilization to nearly 80 percent
• Saved $19,000 per month in power and cooling costs
• Achieved a 12:1 server consolidation ratio
• Realized ROI upon deployment of its second virtual machine—currently have deployed approximately 1,600 virtual machines
• Reduced data center space by a 12:1 ratio
• Virtualized 60 percent of x86 environment
• OS instance doubled while staff remained flat over last three years
• Deploy new servers in hours rather than weeks
• Instituted a VMware-first policy
• Decommissioned more virtual machines than are currently active in efforts to manage virtual server sprawl
• Avoided building at least one data center
• Forewent the following capital expenses: $1.87 million in space, power and cooling infrastructure; $90,000 in wiring infrastructure; and $309,000 in networking infrastructure

Summary

Data centers electricity usage has become a major IT concern that draws government attention and requires effective solutions. VMware Infrastructure reduces data center energy consumption by running applications on fewer physical servers, which in turn reduces the power and cooling requirements. VMware VMotion and VMware DRS technologies also enable effective power management in the data center. Thousands of VMware customers have achieved significant power and cooling consumption reduction in the data center through server consolidation and effective power management. VMware Infrastructure is one of the most impactful solutions to the data center power crisis.

Download Energy Effeciency White Paper