Page 206 -

10: Green IT Case Studies for Universities and a Large Company 171

physical components can become constrained, preventing additional equip-
ment from being installed.
As of 2006, the subject data center contained 61,000 square feet of wall-
to-wall data center space. With the space required for aisles and equipment
maintenance clearance, usable space was reduced to 44,000 square feet. As of
April 2008, 43,000 square feet of the data center was in use, with demand
requests for more than the remaining 1,000 square feet. The UPS unit was
running at 92 percent (2,227 kilowatts of 2,430 kilowatts installed). The
generators were running at 85 percent (4,229 kilowatts of 4,988 kilowatts
installed, noting that additional devices are supported on the generators that
are not supported on the batteries). The chillers were running at 94 percent
(1,175 tons of 1,250 installed).
Something had to be done to address the growth requirements. For this
situation, we return to the IBM five-step process described previously in
this book.

A Five-Step Approach for an Energy-Efficient Data Center

The continuous five-step process, first described in detail in Chapter 2, ptg
“The Basics of Green IT,” is summarized in Figure 10.2. Four out of five of
these steps involve improving the facilities portion of the data center. The
virtualize step involves improving the IT portion of the data center. It’s
ironic that virtualization is the most promising IT technology to affect the
physical data center in such a positive manner. The reason for this is that vir-
tualization allows an IT environment to significantly reduce the amount of
resources being reserved to handle the times when workload peaks. It does so
by allowing multiple workloads to share resources, including the resources
reserved for growth. Typical distributed server virtualization projects plan to
quadruple the equipment utilization, resulting in a 75 percent reduction in
equipment requirements. This translates to freeing up 75 percent of the
power, space, and cooling resources used by distributed servers. IT business
models such as Information Technology Infrastructure Library (ITIL) include
capacity planning for both facilities and IT as updated in ITIL version 3.