Page 188 -
P. 188

9: Green IT Case Studies for Energy Utilities                  153



                  ♦ Six to 10 LPARS (including two VIO LPARS).
                  ♦ Energy savings: The IBM pSeries Web sites give case studies of
                    more than 50 percent energy reduction consolidating stand-alone
                    pSeries servers onto pSeries virtual servers (LPARS).


                Step 4. Evaluate Cooling
                In July 2007, Variable Frequency Drives were installed for all the cooling
             system blowers. There are 10 chillers, and they are powered down when not
             needed (for example, when outside air in the winter is used for cooling).

                Step 5. Manage and Measure
                There is only one meter for the data center building. In addition to the
             data center, the building includes some conference rooms and a few offices. It
             is estimated that more than 80 percent of the electrical energy consumption
             is due to the data center. The IT team plans to use Active Energy Manager
             (AEM) to measure, monitor, and manage the energy components built into
             the servers and other IT systems (see information in the “Next Steps for This
             Energy Utility’s Green Data Centers” section).                                                    ptg
                The main data center energy use versus the IT server and storage capacity
             has been reduced due to server and storage virtualization and other technol-
             ogy (including cooling and lighting, and so on) to increase efficiency. As
             analysis of the energy use indicates, the KWH usage increased over three
             years by two times due to data center consolidation, and such. However, the
             IT team estimated that server capacity over the same period increased 4 to 5
             times. Server capacity increase is not easy to determine, especially with vir-
                                                              ®
             tual servers and when there is a mix of x86, UNIX , and mainframe
             machines. SPEC Marks discussed in Chapter 7, “The Need for Standard IT
             Energy-Use Metrics,” do provide some comparison metrics, and this is an
             area to be further explored. Chapter 7 also discusses the new SPEC informa-
             tion being gathered on server power and performance, which will prove to be
             useful for building green data centers.
                KWH per month energy consumption at the main data center increased
             from an average of 217K KWH in 2005 to an average of 555K KWH in
             2008 (through 11/08). Thus, the average KWH per month at the main data
             center increased approximately 2.5 times from 2005 through 2008. Server
             capacity during that period increased by 4 to 5 times. There was a noticeable
             peak in the KWH graph during September 2006, when the two large main-
             frames were moved from the Ohio, data center to the main data center as the
             final part of a data center consolidation project.
   183   184   185   186   187   188   189   190   191   192   193