Energy Overhead of the GUI in

ServerOperating SystemsHeather Brotherton

Introduction

This study will

Make a case for reducing use of the graphical user interface

Avoid focus on a particular brand of operating system

OPERATING SYSTEM

Nearly every server, controlling systems resources

Allows us to take advantage of the Cascade Effect

Cascade Effect

Potential Savings

Source: EXP Critical Facilities Inc., Intel Corp.

Example

If a PCI card such as video card were removed for a savings of 41watts from 500 servers in a data center, the cumulative watts saved would be 58,220 watts per year.

At an average of ten cents per kilowatt-hour this results in a savings of$51,035.65 per year.

Experiment

Energy readings were collected for a minimum of one hour using the Watts Up? Meter.

Monitoring Tool

Watts up? Pro universal outlet version. This meter is capable of measuring 100 to 250v within a plus or minus 1.5 percent accuracy.

The meter is also capable of logging at one-second intervals and provides a USB interface and PC software

Linux Observations

Linux based server operating systems ran the top command during the observations.

top -d 1 > /home/testOSName.txt

Windows Observations

Windows ran the Typeperf command line tool during the observations configured to provide much of the same information as provided by top.

typeperf “\Memory\Available bytes” “\processor(*)\% processortime” “\Process(*)\Thread Count” > testOSName.csv

Hardware

Intel Atom D525 1.8GHz dual core processor

Integrated Intel Graphics Media Accelerator 3150

Gigabit LAN

SD card reader

5 USB connections

Fan-less external power supply

Intel Solid State Drive 80GB 320 Series

PNY 4GB PC3-10666 1.3GHz DDR3 SoDIMM

Server

Baseline watt consumption mean energy consumed is 7.96 watts and the median is 8.70 watts.

After the addition of 4GB RAM to the server during a one hour period is 15.36 watts and the median is 15 watts.

After Solid State Drive (SSD) installation was a mean of 17.42 watts and a median consumption of 17.7 watts.

Baseline for the server of 17.42 to17.7 watts

Server Operating Systems

The software used for the testing were the following x86 operating systems: Ubuntu 11.10 (Linux) Windows Server 2008 R2 Datacenter

Core Windows Server 2008 R2 Datacenter

GUI

Consumption by OS

GUI vs Non GUITime(Minutes)

Threads(Mean)

Watts(Mean)

Time(Minutes)

Threads(Mean)

Watts(Mean) 

9:42 373 18.65 3:16 263 17.1

9:43 361 18.6 3:17 263 17

9:44 362 18.65 3:18 260 17

9:45 360 18.65 3:19 256 17.05

9:46 360 18.7 3:20 259 17.05

9:47 368 18.7 3:21 255 17.15

9:48 364 18.6 3:22 255 17.25

9:49 363 18.7 3:23 252 17.1

9:50 362 18.65 3:24 250 17.2

9:51 364 18.7 3:25 254 17.2

9:52 376 18.65 3:26 250 17.25

Table explained

The mean number of threads: GUI 365 Non-GUI 256

Difference approximately 109 threads

Indicates that a reduction of the ~100-thread GUI overhead can save roughly one watt at the server level.

FINDINGS

Operating systems that do not run a graphical user interface (GUI) tested use roughly 17.5 to 17.6 watts.

Graphical user interface (GUI) based operating systems tested consumed 18.1 to 18.9 watts roughly.

Not using a GUI would save .6 to 1.3 watts per server.

Conclusion

Savings of roughly 1 watt per server

Doesn’t seem like a big deal?

Maybe, but now you don’t need that video card…

Math

(1watt GUI + 41 watt video card) 2.84 Cascade Effect = 119.28 watts

Hours in a year 8765.81

wattage   x   hours used  ÷  1000  x  price per kWh  =   cost of electricity

(119.28 x 8765.81 ÷ 1000) x .1 = 104.55858168

For 500 servers $52,279.29