Space Weather: A Brief Overview - 7x24 Exchange …Syska Hennessy Group Administrative Director Kathleen A. Dolci 646.486.3818 x103 Membership & Education Tara Oehlmann, Ed.M. 646.486.3818

T H E E N D - T O - E N D R E L I A B I L I T Y F O R U M

SPRING 2011 • www.7x24exchange.org

PUBLISHED BY 7X24 EXCHANGE INTERNATIONAL

Robert F. Kennedy, Jr.2011 SPRING CONFERENCEKEYNOTE SPEAKER

A New Data Center ThatWill Hit Your Sweet Spot

Duffy Development Services

PUE VS PUEHP Technology Services Strategy

Space Weather:A Brief Overview

PUE VS PUE 6

PERFORMANCE MEASUREMENT - A CALL TO ACTION 18

SPACE WEATHER: A BRIEF OVERVIEW 22

HIGH EFFICIENCY UPS SYSTEMS FOR A POWER HUNGRY WORLD 28

TRANSFORMERLESS UPS SYSTEMS AND THE 9900 38

WHAT YOU SHOULD KNOW ABOUT PROTECTING YOUR ELECTRONIC APPLICATIONS FROM DAMAGING FIRES 54

A NEW DATA CENTER THAT WILL HIT YOUR SWEET SPOT 58

THE GREEN REVOLUTION: KEEPING YOUR DATA CENTERS AT MAXIMUM EFFICIENCY 66

INSIDE 7x24 72

3

WWW.7x24EXCHANGE.ORG

Directors and Officers

Chairman of the Board

Robert J. CassilianoBusiness Information Services, Inc.

President

William LeedeckeVanguard

Vice President

David SchirmacherFieldview Solutions

Director – Vendor Representative

Juli IerulliCaterpillar

Director

Cyrus J. Izzo, P.E.Syska Hennessy Group

Administrative Director

Kathleen A. Dolci646.486.3818 x103

Membership & Education

Tara Oehlmann, Ed.M.646.486.3818 x104

Conferences

Brandon A. Dolci, CMP646.486.3818 x108

6658

www.7x24exchange.org

628

CONTENTS

4

7x24MAGAZINE SPRING 2011

After “Digging Out” all winter I am sure everyone is looking forward to Spring!

Although economic uncertainty continues to prevail, 2011 is showing promise of increasedbusiness activity. Companies are slowly but steadily increasing hiring of employees andconsultants to support business needs. Additionally, projects on hold and new initiatives are nowgetting reviewed for approval and implementation. These signs are very positive indicators forMission Critical professionals as upgrading of information infrastructures and building newfacilities are on the list of projects to be approved. A compelling business case needs to be madeto senior management, however the time is right as capacity constraints and aging infrastructuresbecome the driving factors. Companies supporting the Mission Critical Industry that areprepared to very quickly provide knowledge and resources as clients request will be successful andthose that are slow to react will have missed opportunity.

The theme for the 2011 7x24 Exchange Spring Conference being held at the Hilton BonnetCreek in Orlando, Florida June 12 – 15, 2011 is End to End Reliability: “Emerging Trends”.Conference highlights are as follows:

• Conference Keynote: “Green Gold Rush: A Vision for Energy Independence, Jobs andNational Wealth” presented by Robert F. Kennedy, Jr.

• Keynotes by Yahoo and Wells Fargo

• US Department of Energy (DOE) Data Center Practitioner Training & Certification

• Special presentation by NASA on Solar Flares

• Presentations by The Green Grid and the Uptime Institute

• Exchange Tables on specific topics at Monday breakfast and Tuesday lunch

• An End-User Exchange Forum Luncheon on Monday

• Vendor Knowledge Exchange on Monday afternoon

Sponsored Event: “An Evening at Sea World Orlando”

The program content is designed to provide value to conference participants and their companiesby focusing on important topics of the day. Energy Efficiency, Data Center Design, andReliability are highlighted at this year’s Spring event.

I look forward to seeing you at our SpringConference in Orlando, Florida!

Sincerely,

Robert J. Cassiliano

CHAIRMAN’S LETTER

7x24 Exchange Chairman, Bob Cassiliano, presents the keynotespeaker, NFL Legend Joe Theismann with a donation to St.Jude Children’s Research Hospital on his behalf.

6

Reliability, resiliency, energy efficiencyand sustainability are just a few of themany criteria used to judge theeffectiveness of a data center. Now thatenergy efficiency strategies and metricssuch as PUE/DCiE have been vetted foruse in designing, monitoring andreporting, and other metrics such ascarbon and water use effectiveness areright around the corner, the industry isnow attempting to measure and publishdata center energy consumption in aconsistent and uniform manner. Toenable this, organizations are working onrevising or publishing new standards andguidelines on how, where and at whatintervals readings should be taken.

Current Industry Resourcesand AnalyticsRecommendations for Measuring and Reporting Overall Data CenterEfficiency, unites many unrelatedguidelines currently being used in theindustry for the measurement of datacenter energy consumption. Thisdocument is unique in that it has beendeveloped by an industry consortium:

7x24 Exchange, ASHRAE, The GreenGrid, Silicon Valley Leadership Group,Energy Department’s Save Energy Nowand Federal Energy ManagementPrograms, the Environmental ProtectionAgency’s Energy Star program, the U.S.Green Building Council, and the UptimeInstitute all participated. Other seminalprograms include the US Green BuildingCouncil’s LEED Data CenterAdaptations and the EPA’s Energy Starfor Data Centers.

In all of these programs it is essential tounderstand that energy consumption andPUE should not be used as comparativemetrics. It might be intuitive that a facilityin a cold climate will have very differentenergy consumption compared to afacility in a hot climate. But thisdiscussion becomes obscured whenother operational parameters are throwninto the mix – what happens when wemeasure electrical consumption for avery lightly loaded data center? What isthe impact of using different types ofeconomizers (water, direct, indirect,evaporative)? Will a moderate and dryclimate assist in lowering energy use? Or will a moderate and humid climate be

better? How does an off-line UPScompare with a double-conversion UPS?

Now take these options and put theminto a matrix to compare combinations ofalternatives. The matrix gets very large,very quickly. Since there are so manynuances and subtleties in data centerdesign and operation, it becomesobvious that comparing PUEs ofdifferent facilities becomesimpracticable. However, it is possible touse these tools for data center designand benchmarking of existing facilities;tools that take into account all of thesubtleties and create a platform forparametric analysis of the myriadoptions.

Geography and ClimateThe location of the data center impactsthe PUE and annual electricity usage.This primarily comes from the hourlyoutside dry bulb temperature andmoisture content levels that vary basedon location. During the preliminarydesign phase of a project, high-leveldecisions can be made by analyzinghourly weather data. This data includes


by Bill Kosik, PE, CEM, LEED AP, BEMP

8


weather station number, latitude,longitude, elevation, hourly dry bulb anddew point temperatures, and relativehumidities. Using standard psychrometricalgorithms, additional information suchas wet bulb temperature, enthalpy andmoisture content can be generated foruse in more detailed analysis. This datacan also be used to create an annualprofile of dry bulb and dew pointtemperatures. This data visualizationtechnique helps both designers and end-users understand the basics oftemperatures that can be expected in aparticular climate, what the peak valuesare and when they occur, as well as theannual variation in temperatures.(Figures 1, 2)

Building Envelope HeatTransfer and LeakageMaintaining the required temperatureand humidity tolerances in a data centerfor 8760 hours per year is very energy-intensive. Rightfully so, much attentionand research is currently aimed at HVACsystem control strategies and systemefficiencies to reduce energy usage. Onearea that is not being fully addressed ishow the building that houses thecomputers affects the temperature,humidity and energy use. Buildingleakage will impact the internaltemperature and moisture content byoutside air infiltration and moisturemigration. Depending on the climate,building leakage can negativelyimpact both the energy use of thefacility as well as the indoor moisturecontent of the air. Interestingly,depending on the climate, buildingleakage may actually reduce energyconsumption by providingsupplemental cooling and/orhumidification/dehumidification.(Using uncontrolled outside airinfiltration, however, is not arecommended way ofrelying on outside air toreduce the mechanicalcooling load).

Based on a number ofstudies from NIST, CIBSEand ASHRAE investigatingleakage in building envelope

components, it is clear that often timesbuilding leakage is underestimated by asignificant amount. Also, there is not aconsistent standard on which to basebuilding air leakage. For example:

• CIBSE TM-23, Testing Buildings forAir Leakage and the Air TightnessTesting and Measurement Association(ATTMA) TS1 recommend building airleakage rates from 0.11 to 0.33CFM/ft2.

Figure 1 – Knowing the climate in which the data center is to be located is critical in the cooling system selection process.

Figure 2 – Each cooling system will require a different amount of mechanical cooling based on air setpoints and climate.

10

• Data from Chapter 27, “Ventilation andAir Infiltration” from ASHRAEFundamentals show rates of 0.10, 0.30and 0.60 CFM/ft2 for tight, averageand leaky building envelopes.

• The NIST report of over 300 existingU.S., Canadian and UK buildingsshowed leakage rates ranging from0.47 to 2.7 CFM/ft2 of above-gradebuilding envelope area.

• The ASHRAE Humidity Control DesignGuide indicates typical commercialbuildings have leakage rates of 0.33 to2 air changes per hour and buildingsconstructed in the 1980s and 1990sare not significantly tighter than thoseconstructed in the 1950s, 1960s and1970s.

Cooling System StrategiesThe energy use of the HVAC system(DX, chilled-water, air-cooled, water-cooled) will vary based on the outdoor airdry-bulb and wet-bulb temperatures, thechilled-water supply temperature (forchilled-water systems only), and the wet-bulb temperature entering the coolingcoil (DX systems). The annual energyconsumption of a particular HVACsystem is determined by the use ofalgorithms developed to estimateelectrical usage of vapor compressioncooling equipment. Depending on thetype of HVAC system being considered,the variables used in these algorithmsrepresent temperatures for outdoor wetbulb, outdoor dry bulb, chilled-watersupply, and condenser water supply.

One important energy reduction strategyspecific to data centers is using high-temperature chilled water. (In othercommercial building types where humancomfort is the primary goal, colder wateris necessary to achieve both indoor dry-bulb and wet-bulb temperatures asrequired by the standards on thermalcomfort). Using higher watertemperature also enables extendedhours per year to use the economizercycle. These savings can be achievedwith both air and water economizer. Thisextension in hours comes from raisingthe air supply temperature from a typical55°F to 70°F or even as high as 80°F.Since the compressor energy is thesecond highest energy consumer in adata center (next to the serversthemselves), it is important to considerthese ideas when designing orretrofitting a data center.

Part Load EfficiencyMost data centers have a prolongedramp-up of IT equipment installation.This means that for a long period of timethe data center will run at a fraction of itsdesign load. Running the compressorsand UPS systems at low loads generallywill result in low efficiencies. To minimizethe impact of low load conditions,designing modular data centers withexpansion capability will maintainequipment efficiencies at optimal levelsthroughout the life of the data center.

At initial move-in, it is typical to see PUEvalues that are an order of magnitudehigher that what the facility wasdesigned to. Why does this happen? Inaddition to the inefficiencies of thepower and cooling systems noted above,the data center relies on HVAC andelectrical infrastructure and othersupport spaces. So the PUE mustaccount for a very large “overhead” for avery small IT load. Since this “overhead”is in the numerator of the PUEcalculation at the IT is in thedenominator, the PUE increasescommensurately. (Figures 3, 4)


Figure 3 – The building that the data center is located in will consume a basic amount of energy annually.This energy can be significant in the early life of a data center relative to the IT load.

Figure 4 – The PUE of the data center can be significantly higher than the design specificationsdepending on the duration of the IT equipment ramp-up.

12


Interplay with IT EquipmentThe fact that the IT equipment is theprimary “tenant” in a data center facilitynecessitates an on-going dialog tounderstand the needs of the “tenant”.This interface arguably presents thegreatest opportunities for overall energyuse optimization in the data center. Froma thermal standpoint the computer’smain mission is to keep its internalcomponents at a prescribed temperatureto minimize risk of thermal shutdown,reduce electrical leakage and in extremecases, mitigate any chances of physicaldamage to the equipment. The goodnews is that thermal engineers for the ITequipment understand the effects ofwide temperature and humidity swingson the equipment and on thecorresponding energy consumption.From a data center cooling perspective,it is essential to understand how theambient temperature affects the poweruse of the computers. Based on the inlettemperature of the computer, the overallsystem power will change; assuming aconstant workload, the server fan powerwill increase as the inlet temperatureincreases. The data center coolingstrategy must account for the operationof the computers to avoid anunintentional increase in energy use byraising the inlet temperature too high.(Figure 5)

Overall Building EnergyConsumption and PUEIn commercial buildings the outsideconditions impact the energy use of thefacility significantly. Passive designtechniques to reduce energyconsumption when choosing glazingtype, insulation, building orientation,thermal mass, etc. are essential inminimizing unwanted heat transferacross the building envelope. Since mostfacilities such as offices, hotels andschool have a cooling load that ispredominantly externally loaded, thefocus tends to be on how the climateaffects the overall energy use. Datacenters are predominantly internallyloaded - the primary energy reductionstrategies have mostly dealt with

improving the efficiency of the powerdistribution and cooling systems byminimizing losses and improving airdistribution effectiveness. Even thoughthe internal load far exceeds the coolingload imposed by the climate conditions,climate impacts data center energy usein ways that are unique to this buildingtype.

The HVAC system energy will changebased on three main areas: outdoorconditions (temperature and humidity),the use of air and water economizers,and (when designing a new facility)using different types of HVAC systemdesign. These items are detailed in thefollowing.

1. The HVAC energy consumption isclosely related to the outdoortemperature and humidity levels. Thisis because the HVAC equipment takesthe heat from the data center andtransfers it outdoors. The higher theoutdoor air temperature is (and thehigher the humidity level is for water-cooled systems) more work is requiredof the compressors to lower the airtemperature back down to therequired levels in the data center.

2. Economization for HVAC systems is aprocess in which the outdoorconditions allow for reducedcompressor power (or even allowingfor complete shutdown of thecompressors). This is achieved bysupplying cool air directly to the datacenter (direct air economizer) or, as inwater-cooled systems, cooling thewater and then using the cool water inplace of chilled water that wouldnormally be created usingcompressors.

3. Different HVAC system types havedifferent levels of energy consumption.And the different types of systems willperform differently in different

climates. As an example, in hot anddry climates water-cooled equipmentgenerally consumes less energy thanair-cooled systems. Conversely incooled climates that have highermoisture levels, air-cooled equipmentwill use less energy. The design andoperation of the systems will alsoimpact energy (possibly the greatestimpact). The supply air temperatureand allowable humidity levels in thedata center will have an influence onthe annual energy consumption.

Figure 5 – An example of how fans in a server will increase airflow (and power) as the inlet temperatureincreases. Every computer will have different characteristics that must be considered in theoverall data center cooling strategy.

14

Using Energy Analytics forMaking DecisionsIn order to determine how each of theseitems influences PUE, it is necessary todevelop a detailed parametric energyanalysis for several climactic locations.These analyses will consider both HVACsystem type and economizer type.Knowing that hourly weather data isavailable for over 2100 locationsworldwide and knowing that it is notpractical to generate such a high numberof individual analyses, a technique calledmultivariate interpolation is used as ameans to perform the analysis with anacceptable degree of accuracy. Thistechnique, often used in agro-climaticstudies, is a process to assign values tounknown points by using values from ascattered set of known points.

The first step in developing the PUEanalysis is to determine the locationsthat will be used in the in theinvestigation of annual energy use. Forthe analysis illustrated here and toensure that the analysis is based on ageographically diverse set of climaticdata, 118 world-wide locations wereselected based on the prevalence ofdata centers in the area.

Energy SimulationEnergy simulation is the foundationalelement determining data center energyconsumption. A subset of the analysis isto investigate how climate affects PUE.In order to understand the degree towhich the different system types areimpacted in climatic variations, theanalysis includes multiple system types:

1. Air-Cooled DX Direct Air Economizer

2. Air-Cooled DX Direct EvaporativeCooling

3. Air-Cooled DX Indirect Air Economizer

4. Air-Cooled DX Indirect EvaporativeCooling

5. Air-Cooled Chiller

6. Air-Cooled Chiller Direct AirEconomizer

7. Air-Cooled Chiller Direct EvaporativeCooling

8. Air-Cooled Chiller Indirect AirEconomizer

9. Air-Cooled Chiller Indirect EvaporativeCooling

10. Air-Cooled Chiller with Free CoolingModule

11. Air-Cooled Chiller with WaterEconomizer and OA Economizer

12. Water-Cooled Chiller

13. Water-Cooled Chiller - Closed CircuitTower

14. Water Cooled Chiller Direct AirEconomizer

15. Water-Cooled Chiller WaterEconomizer

16. Water-Cooled Chiller DirectEvaporative Cooling

17. Water-Cooled Chiller IndirectEvaporative Cooling

The analysis generates hourly energyuse data in kWH for the following sub-systems (equipment varies based onsystem type):

• Information technology equipment(computers, servers, storage,networking gear)

• Cooling equipment (air-cooled andwater-cooled chillers, air-cooledcondensing unit)

• Fans (for CRAHs, AHUs, exhaust fans)

• Pumps (chilled and condenser water)

• Heat rejection (cooling towers, drycoolers, air-cooled condensers)

• Humidification (adiabatic, isothermal)

It is necessary to have this level of detailin the analysis since the energy use ofmany of the sub-systems is dependenton the climate in which the data center islocated. As an example, the powerdemand of compressorized equipment(chillers, condensing units) is directlydependent on the outdoor temperature(dry-bulb or wet-bulb depending on thetype of cooling equipment). Knowing thateach climate will have a uniquetemperature profile over the course of ayear, the power-per-unit-time, or energy,will also be unique. With the annualenergy values for each of these sub-systems, the PUE can then bedetermined. (Figure 6)


Figure 6a - Calculated PUE values for data centers with water-cooled cooling systems at different latitudes.

16


Figure 7 – Visualization of PUE data using multivariate interpolation will play a key role in understanding the effects of climate on data center energy consumption.

Bill Kosik is Principal Data Center Energy Technologist for HP Enterprise Business, Technology Services. He can be reached at [email protected].

Multivariate InterpolationOnce the PUEs are established for thedifferent systems, the values are used ina multivariate interpolation technique todetermine PUE values that are notknown (essentially “filling in the blanks”).

This analysis yields an output thatvisualizes the PUE data; this will assistthe user to more quickly understand theinfluence of climatic variation. (Figure 7)The overarching purpose of this analysisis to determine if climate has an effecton the PUE of a data center. Some of

the systems exhibit modest percentvariations in PUE across the range ofclimates (water-cooled chiller at 3%) butothers show significant variation inenergy use and PUE (air-cooled chillerwith indirect air economizer at 22%).

ConclusionBuilding design is a process with manycontributors from many differentdisciplines. Buildings that house datacenters have opportunities for energysavings, but to take advantage of theseopportunities, there needs to be aseamless interface between the IT andfacilities teams. This interface also needsto extend out to the IT equipmentmanufacturers to understand thecapabilities and limits of operation priorto final design of the power and coolingsystems. This data along with analysis onclimate, building envelope, part loadefficiency and cooling systems willprovide a solid foundation for a robustand efficiency data center.

Figure 6b - Calculated PUE values for data centers with air-cooled cooling systems at different latitudes.

18

We work in an industry where timelyaccurate data is king, where processesand reports are delivered instantaneouslyand where, when you type a questionyou get an immediate response. So whyis it then, that when we look for specifics,performance metrics and growth trendson our own industry we only findfragmented excerpts, grossly outdateddata and projections based upon non-facts.

Let’s start with the total energyconsumed by data centers. Way back inFebruary of 2007 Jonathan Koomeywrote a research paper revealing that:

1. “Aggregate electricity use for serversdoubled over the period 2000 to 2005…”

2. “Total power used by servers representedabout 0.6% of total U.S. electricityconsumption in 2005. When cooling andauxiliary infrastructure are included, thatnumber grows to 1.2%, …”

When this research was laterincorporated into the August 2007 EPAReport To Congress with the nowinfamous graph below, it energized a ourindustry to develop new and moreefficient ways of consuming energy. Itinspired programs like the Silicon ValleyLeadership’s great “Chill-Off”competitions comparing various coolingtechnologies and continues to be thebasis of many new data center designs.

POWER REPORTINGThis may all seem exceptional exceptthat, what is shown as historical is reallyestimated. Now, I grant you, that to dateit is the best estimate possible but it isstill an estimate. With all our abilities totrack information, with all our computepower, with all our databases and withthe extensive improvements in datacenter power monitoring, it begs thequestion that why four years later we areno closer to getting accurate data thanwhen Jonathan Koomey wrote hisoriginal report? Why, in a world where weget monthly reports on employmentlevels, where we know how many cars

are sold weekly right down to the makesand models; Why can we not even getbasic stats on data center power usage?Note: Koomey only estimated servers notdata centers.

Taking this to the next level, the graphlays out five trend projections, yet in thefour years that have followed, there hasbeen no updating of the data, even on anestimated basis. How are we doing?Which performance curve are wetrending against?

The Koomey research and the EPAreport were at that time enlightening,inspiring and have since had a profound


PERFORMANCE MEASUREMENTA cAll to Action by Dennis Cronin

Comparison of Projected Electricity Use, All Scenarios, 2007 to 2011

10


impact on our industry. Yet, for all thegood that came out of them, without thedevelopment of accurate data to followthere is no way to measure ourperformance as an industry and withoutperformance measurements we will notbe able to manage growth.

The old saying “You cannot manage whatyou cannot measure” was never moreapplicable than it is here.

PERFORMANCEBeyond just measuring raw consumptionwe need to gain insight into how thepower is being consumed, what isconsidered good practice, what is theindustry norm and is the industry gettingbetter or worse?

This then brings us to the muchmaligned acronym “PUE”. (Total datacenter power/IT power). PUE hasestablished what is considered good

practice – get as close to a PUE = 1 aspossible.

Although this is not a perfect metric, ithas established a sustainable industrystandard so long as people do not tryand game the system.

What we are still lacking is anyunderstanding of what the industry normis and is PUE getting better or worse? Iread articles and listen at webinars I hearhow data centers are operating at a PUEof 2.0 or greater yet when I talk tooperators their PUE is always below 1.8and designers are claiming designs of1.4 or better. There seems to be asubstantial disconnect here and it willnever be resolved until we have definitivedata from the majority of the industry.

With such data we could then gaininsights such as year to yearimprovements and are there substantialdifferences between Collocation Vs.Enterprise operations. There are many

things we could learn from having thisdata but we do not have it and as anindustry we are not set up to even begincollecting it.

This is a call to action to all of you asindustry leaders. There are over 1,100collocation facilities in the US alone andat least as many enterprise data centers.With the extensive use of PowerMonitoring and data collection, thesessites at best can analyze their ownperformance but lack comparative datawith which to rank themselves against.We need cooperation, sharing andregular reporting of data. We need acentral repository accessible by all thosewho contribute to the program and weneed professional analysis of the data tohelp guide the future of our industry.

Let’s hear your level of interest in thisventure and together we can make greatthings happen.

Dennis Cronin is the Principal at Gilbane Mission Critical. He can be reached by [email protected].

22


On Earth, we live inside an atmospherethat sustains life but also producesdynamic and often destructivephenomena to which we must adjust ourlives. This includes hurricanes, tornadoes,torrential rains and powerful winds. Anidea that is foreign to many is that wealso live inside the atmosphere of a star,that is the entire solar system sits insidethe atmosphere of a star, an active anddynamic star we call the Sun. Thisatmosphere consisting of ionized gas(plasma), particles and magnetic fieldinteracts with the interplanetaryenvironment producing different effectsthat we collectively call space weather.On Earth and in our local space aroundthe Earth this space weather canproduce beautiful (northern lights) anddestructive results (power andcommunication blackouts). Here we willa brief overview of the type ofphenomena that make up spaceweather, the effects we can experience

on and near the Earth and some of theindustries directly affected by it.

THE SUN HAS A CYCLE

Before we discuss the specifics of spaceweather it is important to mention thesolar activity cycle or solar cycle. Thetypes of solar activity that drive andmake up what we call space weatherfollow a cycle of intensity that is roughlyan 11-year cycle. A traditional measureof this activity is the number of sunspotsvisible on the Earthward side of the Sun.Sunspots are regions of intensemagnetic field that suppress the flow ofheat and light making them cooler thansurrounding areas thus appearing darker.The solar cycle starts off at its minimumwith few visible sunspots and low (calledthe quiet sun) activity. During this timethe solar system is bathed in a steadyflow of matter and energy mostly fromthe solar poles and cosmic rays from

outside our solar system. The steadyflow off the sun is the extended solaratmosphere called the solar wind. Thiswind is composed of energetic particlesof mostly protons and electrons as wellas magnetic fields. The solar wind andcosmic rays produce space weathereffects that though generally smallerthan effects during more active periodsare still dangerous. As the solar cycleapproaches its maximum, the number ofvisible sunspots increases as does thenumber of violent releases of energy andmatter producing more intense magneticand radiation storms at Earth. Figure 1shows the sunspot number since 1700.The black curves are composed fromyearly numbers and the red curves frommonthly numbers.

Our current cycle (cycle 24) started in2008. Figure 2 shows data from theprevious solar cycle (23) along with thecurrent cycle (24) up to April 2011. Also

Figure 1 - Sunspot number from 1700 to the present. Black denotes yearly numbers and red denotes monthly numbers.

by C. Alex Young, Ph.D.

24


shown is the prediction for the remainder of cycle24. The next solar maximum is estimated to be in2013-2014.

THE SPACE WEATHER ZOO

The regions overlying sunspots are called activeregions. Here the sun’s magnetic field becomesconcentrated and twisted because of the motionsof the solar atmosphere at and below the solarsurface. As these regions become more complexthey can eventually become unstable causing therelease of the magnetic energy. This is analogousto twisting a rubber band tighter and tighter untilit snaps releasing energy in the form of heat andmotion. The same thing happens in the solaratmosphere with the active region magneticfields. This release of energy heats up andaccelerates solar material. One form of thisenergy release is a flash of electromagneticradiation over the entire spectrum (radio togamma-rays) called a solar flare. Also during thisenergy release a high-speed blast of material canbe ejected from the sun called a coronal massejection or CME. Both solar flares and CMEscan accelerate particles, mostly electrons andprotons, to extremely high speeds close to thespeed of light, creating a radiation storm called asolar energetic particle event or SEP. Solarflares and CMEs can each occur separately butthe largest (and most dangerous) solar eruptionshave both solar flares and CMEs associated withthem. SEPs generally only occur in the largestsolar eruptive events. Figure 3 shows an example

of a very large solar eruptive event from thefamous 2 week series of space weather stormsfrom October-November 2003. This events arereferred to as the Halloween Storms.

SPACE WEATHER EFFECTS

The solar wind fills the solar system and theEarth is constantly bombarded by this steadyflow of particles and magnetic field. Normally,the Earth is well shielded from this matter by itsself-generated magnetic field. Sometimesbursts of higher speed and density solar windare able to create a disturbance in the Earth’smagnetic field and these disturbances cancreate currents in the upper atmosphere thatexcite atoms creating the northern and southernlights. This is called a geomagnetic storm.

CMEs are the most dangerous and destructivespace weather events. When the Sun producesCMEs that travel to the Earth they can carrymagnetic field of opposite polarity to that of theEarth. This in combination with a solar windenhanced by the CME weakens the Earth’smagnetic field allowing energy to reach theEarth. This creates rapid changes in the Earth’s

Figure 2 - Data for sunspot cycles 23 and 24 up to April 2011. Included is a prediction for the remainderof cycle 24, which estimates the next solar maximum around 2013-2014.

Figure 3 - On the morning of Tuesday October 28, 2003 a large sunspot group, labelled NOAAActive region 10486 (dark patch on the bottom right of the solar disk seen in the image in the upperleft), erupted around 11:12 UT with a fast CME (observed first in the lower left SOHO LASCOimage) and the second largest solar flare observed from space (seen in the upper right SOHO EITimage). The flare and CME produced a strong high-energy particle event that was observed hittingthe SOHO spacecraft 1-2 hours later (creating the “snow” in the lower right LASCO image). ThisCME triggered powerful magnetic storms that caused problems for the electric grid and in-orbitsatellite anomalies and failures. credit: NASA/SOHO

26

magnetic field causing a strong magneticstorm. These magnetic storms inducecurrents on the ground in power grids,pipelines, railway signaling and magneticfluctuations that effect surveying anddrilling for oil and gas.

Another consequence of magneticstorms is heating of the upperatmosphere. This changes the density inthe ionosphere disrupting radio signals,in particular satellite navigation systems.

Solar flares, though visually dramatic andfirst to effect us (light takes 8 minutes totravel from the Sun to the Earth), have afew limited effects on radio basedsystems. Solar flares produce a 10-20minute change to the atmosphere thatabsorbs high frequency (HF) radiowaves producing an HF radio blackouton the Sun facing side of the Earth. Also,this change to the atmosphere slowsdown GPS signals producing locationerrors of many meters in GPS receivers.

SEPs created by solar flares and CMEsare very energetic bursts of particles thatpose a major threat to the electronicsand power systems of satellites. TheEarth’s atmosphere is normally a verygood shield against energetic particlesbut during the strongest of these eventsparticles can disrupt electronics inairplanes and even electronics on theground. These particles can pose asignificant health risk to airline crew andpassengers. Particle storms can createthe same atmospheric absorption effectsas solar flares causing HF blackoutsnear the poles.

Solar flares and CMEs can create burstsof radio waves aptly called solar radiobursts. These intense bursts of radio can

potentially disrupt wireless and short-range devices so they are now ofgrowing concern as the use of thistechnology increases (e.g. Wifi).

INDUSTRIES AT RISK

The effects of space weather are ofhuge concern to power and satellitecompanies, the airline and transportationindustries as well as oil and gas pipelineand survey/drilling companies. Even thefinancial sector is becoming aware ofspace weather’s impact. The awarenessof these industries to space weathervaries with the satellite and powerindustries most aware because of theirprevious exposure to its effects.

EXAMPLES AND COST OFMAJOR SPACE WEATHEREVENTS

• September 1859 - Aurora brightenough to read by seen extremely farsouth (Cuba). Telegraph systems ranwithout being powered. Some caughton fire and operators wereelectrocuted. Considered to be thelargest solar storm on record(estimates for an event of thismagnitude today are up to $2 trillion).

• May 1921 - Entire New York CentralRail switching system knocked out.Telegraph and telephone disruptionsworldwide.

• March 1989 - Huge magnetic stormcaused the Quebec power grid to fail in90 seconds leaving millions withoutpower for up to 9 hours (total cost notjust to the Quebec power gridestimated at over $2 billion).

• October/November 2003 - Major

loss of GPS accuracy with errors over50 meters, transformer failures inSouth Africa, major HF radio blackouts,30 satellite anomalies including onesatellite loss (many billions of $s).

CONCLUSIONS

The effects of space weather events andtheir effects on modern technologies arewell understood and well documented.This does not make them any lessimportant because the broader impacton our socioeconomic infrastructure isless understood. Our society and itsdependence upon modern technologiesis constantly evolving and becomingmore complex and interdependent.Though we have experience and havelearned from many episodes of spaceweather over the years we have yet tosee the impact of major space weatherevents upon our current technologicalsociety. The current state of technologyin our society has evolved greatly sincethe last solar maximum. The next solarmaximum is several years away and theeffects of space weather will becomemore apparent as activity begins toincrease. At the same time ouradvancing society will increase ourvulnerability to space weather. Thecurrent solar cycle is not expected to beas intense as recent ones in the past butthis does not mean we can afford becomplacent. In fact, the largest spaceweather event in modern times (perhapsin the last 5000 years) occurred inSeptember 1859 during a low activityperiod of a relatively small solar cycle.The good news is that the most severeeffects of space weather can bemitigated with proper precautions andpreparations.


REFERENCES AND RESOURCESSevere Space Weather Events - Understanding Societal and Economic Impacts Workshop Report, http://www.nap.edu/catalog/12507.htmlSolar and Heliospheric Observatory, SOHO, a NASA/ESA Mission, http://sohowww.nascom.nasa.govSolar Dynamics Observatory, SDO, a NASA Living with a Star Mission, http://sdo.gsfc.nasa.govSolar Terrestrial Relations Observatory, STEREO, a NASA Solar Terrestrial Probes Mission, http://stereo.gsfc.nasa.gov Public Outreach Sites, http:www.spaceweather.com and http://www.thesuntoday.orgNOAA/National Weather Service, Space Weather Prediction Center, http://www.swpc.noaa.gov/Royal Observatory of Belgium, Solar Influences Data Analysis Center, http://sidc.oma.be/

C. Alex Young, Ph.D. is Solar Astrophysicist at NASA's Goddard Space Flight Center with ADNET Systems Inc. and the SOHO/STEREO science team.He can be reached at [email protected].

UPS systems provide power conditioningand backup power to mission criticalfacilities such as datacenters, broadcastsites and hospitals. UPS systems protectthese sites from voltage fluctuationssuch as surges and sags or frequencyfluctuations and also provide ride-through or temporary power to bridgethe gap between a power outage andthe restoration of utility power or thetransfer to a backup generator. The UPSuses a form of short-term (seconds tominutes) energy storage to assist inpower conditioning and power bridgingin the event of a complete outage. Themost common and practical DC energystorage forms are chemical batteries (i.e.,lead acid, NiCd, NiMH, etc.), flywheelsand ultra-capacitors. To perform its twofunctions, a UPS requires energy – inthis case electricity.

The industry measures UPS efficiencyas power out divided by power in withthe UPS consuming a portion of theinput power. The amount of energyconsumed by the UPS representsenergy lost or inefficiency. UPSinefficiency can waste 10 percent ormore of utility input within the UPS itselfand is a significant concern fordatacenter operators, utilities and policymakers. UPS inefficiency amounts tohundreds of thousands of kilowatt hoursper year wasted in the process ofprotecting even a medium sized missioncritical load.

More efficient UPS systems such asActive Power’s CleanSource UPS can

help reduce electrical waste and cost.Proven in the lab to reach at least 96percent efficiency at loads as low as 40percent, CleanSource UPS can reducedatacenter energy losses by multiplemegawatts and hundreds of thousandsof dollars annually compared to doubleconversion UPS systems, while meetingor exceeding the power quality andsystem reliability of other topologies.

UPS EFFICIENCY DEFINED

The efficiency of a UPS, as defined bythe International ElectrotechnicalCommittee, is “the ratio of (active) outputpower to (active) input power underdefined operating conditions,” wheredefined operating conditions refer to aspecific percent load and load type(linear/resistive versus non-linear).1

Active power is measured in watts orkilowatts.

IMPORTANCE OF UPSEFFICIENCY

The power demands of datacenters aresignificant and growing. The U.S.Environmental Protection Agency (EPA)estimated datacenters consumed 61billion kilowatt-hours (kWh) in 2006 at atotal electricity cost of approximately$4.5 billion. The EPA’s baseline forecastpredicts a near doubling of energyconsumption by 2011 to more than 100billion kWh and $7.4 billion, assumingcurrent growth and efficiency trends.2 Inits alternative forecast views, EPAidentified adoption of higher efficiencyUPS systems as a key factor in reducingdatacenter power consumption.3 A studyby Intel Corp. showed typical UPSsystems as contributing 6-7 percentlosses to overall datacenter energy use.4

See Figure 1.


Learn why energy efficiency for uninterruptible power supply (UPS) equipment is soimportant, what can influence UPS energy efficiency and how it can directly impactenergy costs. This paper will compare the differences in energy efficiencies amongvarious UPS topologies available today through laboratory testing. These studiesdemonstrate flywheel based UPS technology achieves much higher efficiency ratingsas compared to legacy battery based UPS systems, resulting in substantial costsavings over the life of the system.

Fig. 1) Sources of Datacenter Power Use

by Todd Kiehn

28

30


At the individual datacenter level,improving UPS system efficiency offersdirect, 24-hour-a-day energy savings,both within the UPS itself and indirectlythrough lower heat loads and evenreduced building transformer losses.When a full datacenter equipment load isserved through a UPS system, even asmall improvement in system efficiency

can yield large annual cost savings.Pacific Gas & Electric (PG&E) estimatesa 15,000 square foot datacenter with ITequipment operating at 50 W per squarefoot requires 6.9 MWh of energyannually for the IT equipment. If the UPSsystem supplying that power has itsefficiency improved by just 5 percentagepoints, the annual energy bill will bereduced by 384,000 kWh, or about$38,000 at $0.10 / kWh, plus significantadditional savings from the reducedcooling load.5

COOLING REQUIREMENTS AS AFUNCTION OF EFFICIENCY

When evaluating a UPS and itsefficiency, it is important to keep in mindthe first law of thermodynamics that“energy can neither be created nordestroyed.” With respect to UPSsystems, the difference in active inputand output power represents heat lossas a result of the activity the UPSperforms. Heat interferes with theenvironmental conditions in a definedspace such as an electrical room and willultimately drive the temperature up andpotentially cause short- or long-termdamage to equipment as it exceedsdesigned temperature thresholds.6 Asound design of an electrical room

includes an air-conditioning or precisioncooling system to maintain a certaintemperature band. The lower theefficiency of the UPS, the more heat isgenerated and the more cooling isrequired in the room, driving up capitalcosts and the ongoing operationalexpenses of the cooling system. As ageneral rule of thumb for an efficient

centralized cooling plant, it takes 0.33kW of power to cool 1 kW of heat. As aresult, the operational cost of a lowefficiency UPS is further exacerbated bythe additional cooling cost.

FACTORS IMPACTING UPSEFFICIENCY

There are two key factors influencingUPS system efficiency: the topology ofthe UPS system itself and the design ofthe datacenter’s power supply anddistribution, which determines the loadfactor of the UPS.

UPS TOPOLOGY

The design of the UPS system itself hasa significant impact on efficiency. Putsimply, some UPS designs are inherentlymore efficient than others. There are twomajor topologies in use today in missioncritical facilities – parallel online (alsoknown as line interactive) and doubleconversion.

PARALLEL ONLINE

Parallel online UPS systems place theinverter and charger circuitry ortransformers in parallel with the AC utilitysignal. This design allows a parallelonline UPS to compensate for over- orunder-voltages in the incoming utilitypower and, with the right electronics, toeliminate transients, voltage fluctuationsor other disturbances. When utility poweris unavailable or reaches unacceptablelimits, a parallel online UPS enters storedenergy mode. The UPS disconnects theload from utility power and reroutes thisload with a static switch to backuppower, provided by a battery or flywheelthrough the inverter.

DOUBLE CONVERSION

Double conversion UPS systemscompletely isolate IT loads fromunconditioned utility power. They receivetheir name, predictably, because theyconvert unconditioned utility power twotimes under normal operating conditions– first from AC to DC electricity and thenback again from DC electricity into ahighly conditioned AC signal. Double

Fig. 3) Double Conversion UPS Architecture

Fig. 2) Parallel Online Architecture

32

conversion UPS systems always providethe load with a conditioned AC signaleven during normal operation when utilitypower is available and no disturbancesare present.

POWER CONDITIONING

To properly understand the differencesbetween double conversion and parallelonline topology, we first have tounderstand what UPS systems areprotecting against.

The Institute of Electrical and ElectronicsEngineers (IEEE) Standard 1159-1995defines seven categories of powerquality disturbances, such as transients,sags, swells and frequency variations.Depending on the quality of the powerbeing delivered to the facility by theutility, one or more of these disturbancesmay occur frequently. At a minimum, thepower conditioning function of the UPSensures output power transmitted to theIT load is well within the tolerance of theIT power supplies.7 It does this bycorrecting utility power qualitydisturbances and delivering conditionedpower to the critical load.

With a double conversion UPS, all poweris rectified from AC to DC and invertedfrom DC to AC, ensuring a perfect sinewave and frequency protection on theoutput and protecting against all seventypes of disturbances. This approachboth exceeds the requirements ofmodern IT equipment power supplies –which do not always require a perfectlyconditioned load – and uses a significantamount of energy.

The parallel online UPS offered byActive Power address most disturbancetypes in an almost identical fashion to adouble conversion UPS. Active Power’sCleanSource UPS uses a fast actingmicroprocessor and fast voltagetransient discharge window whichsamples the incoming 16.67 msec sinewave 333 times every 50 microsecondsand mitigates any transients with outputfilter capacitors. Reviewing the simplifiedone-line diagram for the parallel onlineflywheel UPS, it is apparent the flywheeloutput electrical path is almost identicalto the double conversion UPS. (CompareFigures 2 and 3.)

The most common objection to theparallel online UPS design is frequencyregulation because it is mains-synchronized: whatever the frequency onthe input is what the frequency on theoutput will be. Active Power addressesthis using the same microprocessor, setto allow as little as +/- 0.2 percentfrequency fluctuation to pass through byusing the flywheel to compensate duringthese events.

Ultimately, the parallel online topology isa simpler design with fewer componentsthat is inherently more efficient whileproviding the same protection to themission critical load on the output.Double conversion UPS systemscondition more than required by today’sIT equipment at the expense ofefficiency. As shown below, flywheelbased parallel online UPS systems havebeen proven in laboratory tests to havematerially higher energy efficiencies thandouble conversion systems across allload factors.

In addition to the higher efficiency of theUPS topology, flywheel systems use lessenergy maintaining temperature andother atmospheric conditions thanconventional double conversion UPSwith batteries. The CleanSource UPSavailable from Active Power supports awide temperature operating range (32 –104° Fahrenheit), so it does not need tobe housed in a temperature controlledbattery room. And it takes up less thanhalf the space of an equivalent rated setof batteries. Combined, this results in asignificant reduction in cooling energyrequirements compared to battery UPSsystems.

ECO-MODE

In recent years, UPS vendors havedeveloped an alternative operatingmodel, known as “eco-mode” or “softmode,” for their double conversionsystems as a way to improve efficiency.In eco-mode, a standard doubleconversion UPS keeps only the invertermodule on, while the active input poweris routed through its bypass circuit, notthe main electrical path originallyintended for full power conditioning andbackup. Only in the event of an outage,

momentary or extended, will the UPSswitch the input power from bypass tothe rectifier and inverter. Vendorspromoting this mode claim theirefficiency at 98 to 99 percent across awide range of loads.

Logically, a UPS running in this “eco-mode” would reduce its losses quitesignificantly given it is performing lesswork. But it achieves high efficiency onlyby almost completely abandoning itspower conditioning function. A UPSoperating in eco-mode protects againstonly a full interruption of utility input. Theother six IEEE defined powerdisturbances are not corrected becauseno power conditioning is taking place.Through active sensing of the bypasspath, the UPS will sample the activeinput power, but will generally feed mostof the power straight through, providedthere are no interruptions. In the event ofan interruption to the UPS input, it will,through fast switching, switch from thebypass path to supplying power directlyfrom its DC energy storage (usuallychemical batteries) through the inverter.But no other disturbances are addressedor remedied, so the quality of the activeoutput power is dramatically reduced,leaving the mission critical load exposedand potentially causing long-termdamage of the protected load.

An often overlooked aspect of theviability of eco-mode is its reliability in ademand situation, where demand isdefined as an interruption of theincoming utility. Presumably, the UPSwould switch from bypass mode to DCenergy storage, providing power in anarrow band of time. However, it iscommonly understood that an electricalswitching device represents the singlemost likely cause of a failure in a UPS.Risk-assessment consulting firmMTechnology, Inc., concluded through theuse of probabilistic risk assessment(PRA) that an automatic transfer switch,for instance, participates in about 95percent of expected system failures as itrepresents a single point of failure.8 Thatis the equivalent mean-time-between-failure (MTBF) of 100,000 hours for acomplex electromechanical component.This switch within an eco-mode UPSrepresents an equivalent single point of


34

failure; applying MTechnology’s analysisraises significant reliability questions ofthe eco-mode UPS approach.

In sum, while a standard doubleconversion UPS wastes too muchenergy to condition power compared toparallel online systems like ActivePower’s, a UPS in eco-mode achievesefficiency equivalent to Active Power’sonly by sacrificing too much functionality.The increased probability of failure andthe absence of power conditioning makea double conversion UPS working in“eco-mode” inferior to parallel onlineUPS topology like Active Power’sCleanSource UPS.

UPS LOAD FACTOR

UPS systems in mission criticalenvironments typically operate between30 percent and 80 percent of ratedcapacity depending on the level ofredundancy designed into the electricalsystem.

In facilities with the highest requiredredundancy, UPS systems are deployedin 2N configurations: the load of thefacility is shared between two UPSsystems so that if one fails, the other willstill supply the datacenter’s critical load.In this configuration, the entire missioncritical load cannot be more than half ofthe UPS capacity installed. Under normaloperating conditions, the UPS equipmentnever operates at more than 50 percentload. Typically, these designs end upoperating at 30 percent to 40 percent ofUPS capacity because common practicedictates that UPS systems should not berun at 100 percent capacity even inabnormal operating conditions anddatacenters are usually designed withexcess capacity to permit growth overtime.

In an N+1 configuration, one or moreUPS systems work as a spare, providingredundancy to support the full loadshould any one of the UPS systems fail.In an N+1 system having three UPSmodules, the design load cannot exceed66 percent of the installed UPS capacityand under current design standard willusually not exceed 90 percent of thatlevel, or 60 percent of installed capacity,

when the datacenter is fully populated. Ifthe datacenter is not fully populated,loads can range from 40-60 percent.

UPS efficiency varies with the amount ofactive power being supplied to a loadand tends to decrease at partial load.The efficiency drops significantly atlower loads because the UPS generallyhas a base consumption or fixed“overhead” of power use, similar to a carat idle. The base consumption is drivenby its control circuits, air circulation andcharging currents, which remain virtuallyconstant regardless of loading. Powerused in the conditioning of active outputpower is almost directly proportional tothe load protected, so as the loadincreases the base consumption has lessand less of an impact on energyefficiency.

Energy efficiency at various loads forActive Power’s CleanSource UPSsystem and a typical double conversionUPS system with batteries aresummarized in Figure 4, below.

VENDOR CLAIMS – FACT VERSUS FICTION

Many UPS vendors publish an energyefficiency percentage as part of theirproduct documentation. Most vendorspublish a single “up to” efficiency ratingfor their system in ideal conditions,generally 75 to 100 percent of ratedactive output.9 With users and operators

being very energy conscious in thebuying process, some UPS vendors postfar too optimistic efficiencies of theirUPS systems to attract buyers. This isnot to say the claims are untruthful, butrather they are practically impossible toreplicate in a properly designed missioncritical infrastructure at the end user site.There is a need to be cautious andcritical of the data provided since it canchange dramatically as options (such asfilters and transformers in the case ofdouble conversion UPS with batteries)are added.

Even when vendors provide efficiencyratings across a range of loads, thoseratings often prove to be impossible toreplicate when tested in controlledsituations or measured in the field asdescribed below. By contrast, theflywheel UPS from Active Power hasproven 98 percent efficiency at 100percent loads and 96 percent efficiencyin 2N redundant / 40 percent loads, asshown below.

EMPIRICAL RESULTS:LABORATORY TESTS

A study by Lawrence Berkeley NationalLaboratory (LBNL) examined the energyefficiency of the different UPStopologies across various loadcondition.10


Fig. 4) UPS Efficiency Compared

37


Flywheel UPS systems like Active Powerdemonstrated efficiencies of 95 percentat 33 percent loads, rising to 98 percentat loads above 50 percent. By contrast,double conversion UPS systems showeda wide range of much lower efficiencies

with most exhibiting 80 to 90 percentefficiency at 33 percent load and 85 to94 percent efficiency at 50 percent load.The most efficient double conversionUPS approached 95 percent only asloads exceeded 75 percent with most

still clustered at 85 to 92 percentefficiency.11

These results led both LBNL and PG&Eto strongly endorse flywheel UPStechnology to improve datacenter energyefficiency.12

CONCLUSION

Datacenter energy consumption is asignificant and growing concern foroperators, utilities and policy makers.Inefficient UPS systems can contributeto this concern with 10 percent or moreof utility input going to electrical wastewithin the UPS itself. Flywheel-basedparallel online UPS systems such asActive Power’s CleanSource UPS can bepart of the solution. Proven in the lab toreach at least 96 percent efficiency atloads as low as 40 percent, CleanSourceUPS can reduce datacenter energylosses by multiple megawatts andhundreds of thousands of dollarsannually compared to double conversionUPS systems, while meeting orexceeding the power quality and systemreliability of other topologies.

1. IEC 62040-3, Testing Procedures for UPS Systems. International Electrotechnical Committee. April 30, 2004, at 52.

2. U.S. EPA, Report to Congress on Server and Datacenter Energy Efficiency, Public Law 109-431, Aug. 2, 2007, at 7.

3. Id. at 53, Table 3-5.

4. M. Ton, B. Fortenberry, and W. Tschudi, Lawrence Berkeley National Laboratories, DC Power for Improved Datacenter Efficiency, March 2008, at 18, Fig. 7(citing Intel Corp.). Available at http://hightech.lbl.gov/documents/DATA_CENTERS/DCDemoFinalReport.pdf

5. PG&E, High Performance Data Centers, Jan. 2006, at 55. http://hightech.lbl.gov/documents/data_centers/06_datacenters-pge.pdf

6. The industry rule of thumb is that every 10°C increase in temperature doubles the component failure rate of IT equipment.

7. These tolerances are defined in the Information Technology Industry Council (ITIC) power acceptability curve.

8. See Active Power White Paper #103, Reliability Assessment of Integrated Flywheel UPS versus Double conversion UPS with Batteries.http://www.activepower.com/wpdownload/index.php?wp=WP_103_Reliability_Assessment.pdf

9. See, e.g., Liebert’s NX and Series 610 products highlights at http://www.liebert.com/product_pages/SecondaryCategory.aspx?id=4&hz=60; Piller APOSTAR APPremium overview at http://www.piller.com/site/static/appremium.asp.

10. Lawrence Berkeley National Laboratory, High Performance Buildings: Data Centers, Uninterruptible Power Supplies, December 2005, at 20, Fig. 17.http://hightech.lbl.gov/documents/UPS/Final_UPS_Report.pdf

11. See also id. at 21, Table 4. Note double conversion UPS systems were tested in standard operating mode only; eco-mode claims were not evaluated.

12. Id. at 21; PG&E, High Performance Data Centers, Jan. 2006, at 53.

Todd Kiehn is the Director of Product Development at Active Power. He can be reached at [email protected].

Fig. 5) Factory Measurements of UPS Efficiency Source: LBNL

38

INTRODUCTION

There is a growing trend in the UPSindustry to create a highly efficient, morelightweight and smaller UPS system.Data Centers across the world areconstantly searching for new ways tomaximize reliable UPS power whileminimizing the UPS and associatedequipment footprint. In addition,operating costs for large data centersalways will be a priority.

In applications which require a smallercapacity UPS (less than 200 kW), trueon-line double conversiontransformerless UPS systems haveemerged as the topology of choice. Inlarger applications, most UPS systemsconsist of a UPS with a transformer, ormultiple smaller UPS power modulesparalleled together to achieve therequired capacity. Most UPSmanufacturers are finding it difficult tocreate a true on-line double conversiontransformerless UPS system larger than200 kW due to factors such as groundfaults, high frequency noise andefficiency.

Transformerless UPS systems utilizeInsulated Gate Bipolar Transistor (IGBT)for all power conversion processes(AC/DC converter, DC/DC chopper andthe DC/AC inverter). IGBTs are muchfaster than the traditional thyristor andcan be controlled by simply toggling anon/off gate signal using a digital signalprocessor and a field programmable gatearray as opposed to waiting for a zerocrossing. When the gate signal isremoved, the IGBT turns off. Thecombination creates a series of pulses tore-shape existing voltages (conversionfrom AC to DC and from DC back toAC).

As with any switching power electronic,the device itself has power losses. Forthe IGBT, the two primary losses are theconduction losses and the switchinglosses. Since the IGBTs are being turnedon and turned off much faster, theswitching losses will increase creating aless efficient system. This is a challengemost manufacturers have been unable toovercome in larger capacity UPSsystems.

For this reason, many UPSmanufacturers will still use thyristors asopposed to IGBTs in the convertersection of the UPS (rectifier section).Although there are many benefits inusing IGBTs in the converter section, adecrease in efficiency prevents this frombeing a preferred option for most inlarger UPS kW rated systems (above200 kW). In the UPS inverter, thebenefits of the IGBT switching speedhave far outweighed the decrease inefficiency.

It is important to note that a transformerserves many purposes in the UPSsystem. Even if the UPS utilizes IGBTsfor switching, a transformer would stillhave a purpose. Other technologies suchas the IGBT switching controls and faultdetection still need to be considered.

This paper will discuss the newtechnologies used in transformerlessUPS systems and the advances in theMitsubishi 9900 series UPS systems.

by John Steele

40


BACKGROUND

Before the advances in UPS controls andthe benefits of the IGBT in the converterand inverter can start to be appreciated,the primary and secondary purposes of thetransformer in traditional UPS systemsmust first be discussed.

CONVERTER SECTION:

The main purpose of the converter sectionof the UPS is to convert the AC utilitypower to DC power. The most popularpower electronics used for this processare the Diode (6 pulse), the thyristor (SCR6 pulse and SCR 12 pulse) and the IGBT.Until recently, the 6 pulse and the 12pulse SCR rectifier have been the mostpopular.

As shown in figure 2, the 12 pulse SCRrectifier uses an isolation transformer incombination with two 6 pulse SCRrectifiers (figure 1 shows a 6 pulse SCRrectifier). The 30 Degree phase shiftprovided by the transformer is the mainpurpose of the isolation transformer. Thetwo 6 pulse SCR rectifiers will alternate tocreate twice as many pulses. The 12 pulseSCR rectifier will produce fewer harmonicsthan the single 6 pulse SCR rectifierallowing for a smaller harmonic filter.

The Diode bridge converter (figure 3) issimilar to the 6 pulse SCR rectifier, exceptthe Diodes are natural commutation(natural on and natural off). The benefitsof the Diode bridge are better efficienciesand lower harmonics compared to atraditional SCR rectifier.

Every rectifier, regardless of thetechnology or power electronics used, willproduce harmonics. For all technologiesbut the IGBT (figure 4), these harmonicsare greater than desired for most electricalsystems, including the backup generator.Therefore, an input filter is required toreduce the harmonics to less than 10%iTHD. As shown in Figure 5, the input filteris comprised of an inductor with a parallelcapacitor. The transformer, in this case,helps add inductance to the line to furtherreduce the harmonics. Figure 6 showsreflected harmonic distortion on inputcurrent waveforms from their respectiverectifier or converter technology.

Figure 1: 6 pulse rectifier with optional transformer

Figure 3: Diode Bridge rectifier with optional transformer

Figure 2: 12 pulse rectifier with optional transformer

43


Finally, the isolation transformer used inthe converter section of the UPS systemwill provide isolation between the UPSmodule and the utility. This isolationhelps protect against DC ground faultscaused by the batteries and will minimizedamage to the upstream distributionsystem due to a separation in thetransformer input and output windings.

Although some manufacturers havestarted using IGBTs in the convertersection, most are still switching theIGBTs based on the traditional SCRlogic. The benefit of using a slowerswitching speed for the IGBT converteris a higher efficiency (less switchingmeans less switching losses). In thiscase, the IGBT converter will stillproduce larger amounts of harmonics,which will require the input harmonicfilter (similar to the harmonic filter for theSCR rectifiers).

The problem with traditional harmonicfilters on the UPS converter is theleading power factor at small loads.When the UPS system is operating at areduced load, the ratio of capacitance inthe input filter to the load becomes verylarge and will produce a leading powerfactor from the UPS. This leading powerfactor can result in generatorcompatibility issues. To eliminate theseissues, an active input filter (switchingharmonic filter capacitors in and out ofthe circuit depending on the load) willneed to be used or the input filter wouldneed to be disconnected both resultingin an increased harmonic content.

DIGITALLY CONTROLLED IGBTCONVERTERS:

IGBT converters offer many benefitsover SCR type rectifiers if appliedcorrectly (see figure 4 above). The IGBTconverter can switch at speeds in thekilo-hertz range as opposed to theslower SCR rectifier, which fires pulsesin the hundreds-hertz range.

The reason the SCR rectifiers can not beswitched faster is because the thyristorsare turned on by a gate signal, but turnedoff by natural commutation (zerocrossing of the AC input sine wave) or bya snubber circuit. If used properly with adigital signal processor and a Field

Figure 4: IGBT Converter with optional transformer

Figure 5: Filters for the rectifier section (Left) and inverter section (Right)

Figure 6: Input current harmonic comparison

44

Programmable Gate Array, the IGBTswitching can be controlled to minimizethe harmonics normally produced by aconverter, thus eliminating the inputharmonic filter.

As stated above, the problem exists withthe UPS efficiency. If the IGBT converteris turning on and off in the kilo-hertzrange, and the IGBT inverter is turningon and off in the kilo-hertz range, theswitching losses will quickly add upcreating a less efficient UPS.

INVERTER SECTION:

Most manufacturers have recognized theimportance of using IGBTs in the UPSinverter. The faster switching capabilitiesof the IGBT, if controlled properly, willresult in a less distorted sine wave andwill offer better response to varioussteps in the output loads and improvedcompatibility with downstream statictransfer switches.

The IGBTs are fired in a series of pulses,which are used to invert the DC from theconverter (or batteries) into a clean sinewave (refer to figure 7). Since the pulsesare not continuous, high frequency noiseis generated from the inverter. The highfrequency noise and harmonic contenton the output side of the UPS is filteredusing an output filter and an outputisolation transformer.

In addition, the isolation transformer onthe output of the UPS provides a meansto re-establish a neutral ground bond forthe downstream distribution.

THE MITSUBISHI 9900 SERIESTRANSFORMERLESSUPS SYSTEMS

As mentioned earlier, there are manydifferent factors that need to beconsidered when designing atransformerless UPS system. Ideally, theUPS should be designed with minimallosses throughout the conversionprocesses (high efficient UPS). This wasone of the design goals for the 9900series UPS. After researching the

available power conversion technologies,it was concluded that the three-leveltopology offered the most benefits whileimproving the reliability of the system.

The three-level topology reduces theswitching losses associated with IGBTs.The result is a more efficient True On-Line Double Conversion UPS systemthat maintains this higher level ofefficiency at loads as low as 10% on theUPS system (reference “The Power ofGreen: Mitsubishi 9900A Series HighEfficiency True On-Line DoubleConversion Uninterruptible PowerSupply (UPS)”).

Adding to the efficiency improvements ofthe 9900 series is the reduction of thefiltering requirements. The input currentharmonics of the 9900 is controlled toless than 3% at 100% load without theuse of an input harmonic filter. Only asmall EMI filter is required for the inputof the UPS. By eliminating the inputharmonic filter, generator compatibilityissues due to harmonics and leadingpower factor are eliminated.

On the output, since the Three-LevelTopology generates pulses that moreclosely resemble a sine-wave (see figure7), the output filter is also minimized. Thecombination creates a highly reliable andefficient True On-Line DoubleConversion UPS system. Reducing the

size of the UPS filters and eliminatingthe transformer will also reduce the sizeand weight of the UPS system andincrease the efficiency.

To further enhance the improvements intechnology, the 9900 series UPSsystems are designed with the newest,most advanced generation of Mitsubishidesigned IGBTs. Most manufacturers arestill using the second and thirdgeneration insulated gate bipolartransistors. Mitsubishi UPS systems areusing fourth and fifth generation IGBTswhich offer better efficiencies, improvedreliability and a lower gate voltage.

Although these are all great benefits forthe customer, the UPS controltechniques and the UPS response todifferent types of faults that can occurmust now be considered.

CONTROL SYSTEM:

Switching the IGBTs faster is of greatbenefit, but more important are thecontrol techniques used for the gatesignal. The Mitsubishi 9900 series UPSutilizes a high speed Digital SignalProcessor (DSP), Application SpecificIntegrated Circuit (ASIC) and FieldProgrammable Gate Array (FPGA) forcontrol.

As shown in the block diagram of thecontrol circuit, Figure 8, the 9900 series


Figure 7: IGBT Switching 2-level (top) and 3-level (Bottom)

46


UPS uses a minor current control loopwith a major voltage control loop toprovide a precise response to variousconditions imposed on the UPS.

Data from the voltage major control loopis collected and evaluated in the digitalsignal processor (DSP). The fieldprogrammable gate array (FPGA) isused to collect information provided bythe DSP in addition to comparing thecurrents from the current minor controlloops. Application specific programmingin the FPGA processes the informationand provides a switching sequence tocompensate for the specific scenariosthe load presents.

The multiple feedback control loopallows the UPS logic to quickly detectchanges in current in addition todetecting deviations to the voltageallowing for more precise control of theIGBT switching. By using the DSP andthe FPGA, which have samplings rategreater than 48 kHz (800 times percycle), the switching sequence of theIGBTs can be manipulated to providecomplete control of the powerconversion processes.

In the event a fault occurs (includingshort circuits) which produces currentsexceeding the current limits of the UPS(input or output), the control circuit will

immediately stop firing the IGBTs andinitiate the opening and closing of thecontactors. On the converter section, theinput contactor will open and the UPSwill operate from the battery source. Onthe inverter, the UPS will initiate atransfer to bypass to allow the bypasscurrent protection to clear the fault.

With minimal voltage distortion, the UPSwill support any currents that do notexceed the limits of the system, includingdownstream inrush currents. Forexample, the UPS inverter can support0% to 100% step loads on the UPSoutput while maintaining a voltage withless than 2% deviation (See Figure 9).Since the converter section is alsocontrolled using the same logic, theconverter can support the same steploads to the input of the inverter.Therefore, the UPS does not requirepower from the battery system toperform this step load.

CONTROL IN MULTI-MODULEAPPLICATIONS:

The 9900 series UPS is capable ofmultiple module configurations (MMS),up to eight kVA matched units in parallel.For all multi-module 9900 series UPSsystems, single modules are usedwithout any changes. The same logic isused in each UPS in the multi-moduleconfiguration. For a multi-module system,redundant Cat 5e cables are used toconnect the DSP and the FPGA of eachUPS together. Therefore, the controllogic will be looking at its own minor andmajor control loops, and also theinformation from the other UPS controlloops, to match the output of each UPSon the critical bus. The result is a multi-module system which caninstantaneously share load whilemaintaining clean, reliable, regulated anduninterrupted voltage on the critical bus(refer to figure 10). Since the 9900series multi-module configuration is usedwith individual single module UPSsystems, the control logic is completelyredundant.

VIRTUAL NEUTRAL:

Line-to-line noise produced by the highfrequency switching inside the UPS

Figure 8: Block diagram of the Mitsubishi Control citcuit for a UPS

Figure 9: 100% step load on a 9900 series UPS without batteries

48

system is easily filtered using the smallinput and output filter of the UPS.However, high frequency line-to-neutralcomponents are not suppressed due toan absence of a common connectionbetween the three phases (neutralconnection).

To filter these components, the 9900series UPS uses a virtual neutral asshown in figure 11, which is created byconnecting the common point of each ofthe filter capacitors to a common point.By virtue of this connection, the commonmode harmonics are passed through thevirtual neutral of the UPS system.

In addition to the input connection, thevirtual neutral is also tied to the commonpoint of the output filter, where thecommon mode harmonics are cancelledby the output of the inverter, as shown inFigure 11. During battery operation, theinput contactor for the UPS system(CB1) is opened and the common modeharmonics are eliminated from theequation.

The potential of the Virtual Neutral isderived from the three phases on theinput. A capacitor is added between thesystem ground and the virtual neutral.This capacitor, under normal conditions,will have minimal potential across theterminals and minimal current, as thepotential of the virtual neutral and thesystem ground is the same.

As shown in Figure 12 (including thefollowing calculations), confirmation canbe made that the output phase voltagereferenced to ground will be the same asthe output phase voltage referenced tothe virtual neutral. The input commonmode harmonics are introduced throughthe Virtual Neutral, but cancelled by theoutput common mode harmonics.

FAULT CONDITIONS:

The 9900 UPS system is monitoringfault conditions in multiple ways (voltagedifferences, current flow and currentlimiters) and will detect these faultsdepending on the installation of thesystem, the configuration of thedistribution system and the conditions ofthe fault.


Figure 10: MMS system with (1) module added to the parallel bus

Figure 11: The 9900 series Virtual Neutral

Figure 12: Virtual Neutral with the output voltage to ground reference calculations

50

7X24MAGAZINE SPRING 2011

As mentioned above, if a fault conditionoccurs on the output of the UPS whichproduces a large amount of current, theUPS system current minor loop, currentlimiter and the FPGA will detect theinstantaneous over-current. The PWMwill then stop firing the IGBTs in theinverter and transfer the system tobypass so the bypass over-currentprotection can clear the fault (Figure 16).In addition, during an output short circuit,the output voltage will try to collapse.The voltage major control loop with theDSP along with the FPGA will maintainthe voltage at first. If the current from theshort circuit extends beyond thecapabilities of the inverter logic tomaintain proper voltage regulation, theDSP and the major control loop willsense a collapse in the output voltageand the system will detect aninstantaneous overload.

The same would apply to a fault on theutility side of the UPS system. Theconverter logic will sense an immediatedrop in voltage and the FPGA will stopfiring the IGBTs in the UPS convertergoing immediately to battery operation.

DC GROUND FAULTCONDITIONS:

If a DC Ground fault occurs as in Figure13, the potential of the system groundwill be different than the potential of theVirtual Neutral. This potential differencewill be detected by the UPS system andviewed as a safety hazard. The UPS willalarm on a DC Ground Fault condition.The circuit is also being monitored by thechopper circuit logic to detect impropervoltages and current levels.

GROUND FAULT DETECTION:

As shown in Figure 14, if the utilitysource for the UPS is not grounded (noneutral-to-ground bond), the presence ofa ground fault will present a potentialdifference between the UPS VirtualNeutral and the system ground.Therefore, the fault will be detected bythe UPS.

As shown if Figure 15, if the utilitysource for the UPS is grounded (neutral-to-ground bond is installed), thepresence of a ground fault will result in a

Figure 13: DC Ground Fault

Figure 14: Ungrounded Phase-to-Ground Fault

Figure 15: Grounded Phase-to-Ground Faults

Figure 16: Phase-to-Phase Faults


second connection to ground and willproduce zero sequence currents. Thesezero sequence currents will be detectedby the UPS.

SUMMARY:

The 9900 series UPS is a highefficiency True On-Line DoubleConversion transformerless UPSdesigned specifically for large powerapplications. It uses an IGBT convertersection with advanced control circuitryto eliminate common mode harmonicsand leading power factor issues withthe upstream generators. The minor andmajor control loop used with the DigitalSignal Processor and the FieldProgrammable Gate Array for the IGBTinverter results in precise control of theoutput voltage for maximumcoordination with downstream PowerDistribution Units and Static TransferSwitches. The Virtual Neutral allows theUPS logic to monitor all types of faultsincluding system ground faults to offermaximum protection for critical loads.

The result is a UPS which can deliverreliable output power to your criticalload with minimal footprint while savingthe customer money through anefficiency of greater than 96% at 100%load. The 9900 is available in sizesranging from 72 kW to 750 kW (SMS)and up to 3750 kW (MMS).

REFERENCES: “The Power of Green:

Mitsubishi 9900A Series High

Efficiency True On-Line Double

Conversion Uninterruptible

Power Supply (UPS)” by Dean

Richards and Junichiro Onishi.

John Steele, EIT is the Engineering Managerat Mitsubishi Electric. He can be reached [email protected].

52

54

From a smoldering circuit board to aswitch gear explosion, the risk ofelectrical fire in data, power generationand telecommunication centers is anever-increasing concern for the ownersand operators of these facilities. Today’sdata center and electronics industriesare looking for a fire suppressionsolution that is safe, effective, cost-efficient and mindful of the environment.There are a variety of fire suppressionsystems on the market today that aremade to protect electrical and electronicapplications, from water mist to gaseousto inert gas technologies, however, thosesystems can present design,performance and maintenancechallenges. This article will review thosedifferent technologies and challenges, aswell as a newer fire suppression hybridsolution that uses both water andnitrogen.

AVOIDING EXCESS WATERDAMAGEEquipment enclosures commonly foundin data center and electronic facilities,present significant fire suppressionchallenges and are considered “localapplication” hazards. Traditional, single-agent water mist systems often need tobe installed directly within an enclosureto suppress a fire. This internalinstallation would require the equipmentto be de-energized before a water mistsystem discharge occurs in order toprevent damage and injury.

If fire suppression is required in a largerarea, where multiple enclosures orpieces of equipment may need to beprotected, this would be considered a“total flooding” application. In thisscenario, the water mist concentrationthat enters an enclosure would not be

great enough to cool and suppress thefire unless the top were open.

In either scenario the equipment insidethe enclosure would be subject toextreme wetting conditions along withpossible damage.

MAINTAINING ROOMINTEGRITYAlternatively, gaseous systems requirehold time concentrations and somesystems require the space to bemodified to provide a room integrity sealso the gaseous agents remain in thehazard area during a fire event.

Where water mist systems are designedto reduce the temperature in the firespace, inert gas systems work to reducethe amount of oxygen available tosustain the fire. These particular systems


by Frank Barstow

56


do not allow for cooling in the space,leaving open the possibility of re-ignitionif room integrity is not maintained for adefined period of time.

Effective maintenance of theconcentration of a suppressing agent isa value of room integrity. Proper roomintegrity for many inert gas systemsrequires significant and potentiallyexpensive custom construction to controlairflow, which can be a challenge forfacility owners and contractors.

KEEP IT GREEN AND EASY TOMAINTAINInert gas suppression systems, designedto reduce the amount of oxygenavailable to sustain a fire, also presentseveral challenges in today’ssustainability-focused marketplace.These systems utilize chemical agentsand halocarbons to achieve firesuppression, raising environmentalconcerns. Therefore, efforts have beenmade to find alternative solutions thatare friendly to the environment andoccupants.

Beyond room integrity and environmentalconcerns, the full life cycle of a systemmust be considered, including the costand complexity of resetting a systemafter a discharge. After a fire event orduring regular maintenance, traditionalagent storage cylinders may need to bereturned to the manufacturer for refillwith proprietary agents or removed fromthe system for weighing, costly in bothlost time and money. Facility owners andoperators who are focused on minimizingfacility downtime are in need of moretime and cost-effective solutions.

NEW SUSTAINABLE, HYBRIDTECHNOLOGYA unique solution for data centers useshybrid technology to deliver a highvelocity, low pressure blend of water and

nitrogen to both cool the hazard areaand remove the oxygen that sustains thefire. Electronics are kept dry and thereare absolutely no toxic agents orchemicals involved. Additionally thehybrid system has been proven effectivefor small, smoldering enclosed fires andlarge, heat-releasing fires in openspaces.

By combining nitrogen and water, ahomogeneous suspension of nitrogenand sub 10-micron water dropletspenetrate through vented typeenclosures to extinguish a fire withoutsignificant water residue. When themixture enters the enclosure, both thenitrogen and water attack the firesimultaneously, the water cooling thespace and the nitrogen reducing theoxygen content and generating steam.The result of this is complete firesuppression without significant wettingor change for re-ignition. This would notbe possible with a traditional water mistsystem.

This innovative technology is 100percent green, utilizing no toxicchemicals and has been identified by theEnvironmental Protection Agency ashaving Global Warming Potential of zero.Since this system discharges with highvelocity under low pressure, the hybridsuspension swirls through the space,overpowering the fire plume and notrequiring room integrity to remaineffective, therefore additionalconstruction or retrofit cost can beavoided.

Also, since this technology only requiresthe use of water and nitrogen, storagebottles can be quickly and cost-effectively refilled on site or a local gassupplier can replace cylinders. This aloneminimizes system and facility downtime,increasing productivity and profit. Thesystem is designed to receive waterpiped in or stored in a portable tank. De-ionized water is recommended for

systems where electrical equipment is inthe hazard space to minimize potentialconductivity.

A PROVEN TECHNOLOGYMany have witnessed the effectivenessof a hybrid technology, including FactoryMutual and Underwriter’s Laboratories.For example, the system was tested todemonstrate the successfulextinguishment of a cable bundle fire.Because this technology can be installedin a total flooding configuration, itremains completely effective at fireextinguishment inside an enclosure. Withemitters installed outside the enclosure,the live circuits do not have directexposure to the water and nitrogensuspension discharged by the system,which greatly minimizes the moisturelevel in the enclosure. Due to the highvelocity distribution of the system, thewater and nitrogen mixture travels intothe enclosure through ventilationopenings, ensuring that circuit boardsand other electronic components are notdamaged due to high levels of moistureand/or excess wetting.

CONCLUSIONHybrid technology provides aninnovative, safe and effective fireprotection solution for installations thatcontain electronic equipment. Water mistsystems that use single fluid technologycannot provide the same level ofprotection as a hybrid technology. Sincea single fluid system will need a nozzleplaced inside of the enclosure toextinguish a fire, it is certain that thepowered equipment would be damageddue to extreme wetting conditions.Facility owners and operators will benefitfrom the hybrid technology’s minimalwetting operation, simple maintenance,environmentally friendly design and rapidreturn to normal operations after systemdischarges.

Frank Barstow, Victaulic Vortex™ Sales Manager for Victaulic. He can be reached at [email protected]

58

EXECUTIVE SUMMARYThe decision to buy, build or lease a datacenter is a daunting challenge thatrequires a unique combination offinancial, IT and real estate skills andresources. You always begin by lookinghard for fatal flaws. A lack of reasonably-priced power, adequate water supply, orquality fiber provider should disqualify apotential data center or data center sitefrom consideration.

The optimal data center offers a low totaloperating cost; on-site, reliable power;multiple fiber providers; scal ability andflexibility; efficient, cost-savingtechnology; LEED Gold and/or EnergyStar certification; a secure, desirablelocation; critical infrastructuremanagement; one party decision-making; and flexible lease/purchaseoptions.

Scalability and a low PUE is critical. Totalcost of ownership may be fine today, butmay be compromised when you need toexpand. Long-term, more often than nota facility with a low PUE and morescalability is a better choice than afacility with marginally lower power cost.As the industry trends quickly towardvirtualization and cloud computing, a“next generation” data center avoids theinefficiencies of partially-used servers,racks, and capacity.

It’s all about finding the “sweet spot”where all the essential factors converge.

The decision to buy, build or lease a datacenter is a daunting challenge. It is oneof the most significant financialdecisions a company is likely to make inany given year, requiring consensusamong financial, IT and real estateplanners, the final choice will have anenormous impact on the company’sfuture success.

The right data center should providereliable functionality, security, flexibility,and convenience. It should also adapt asthe future unfolds. The decisiondemands in-depth knowledge of

sophisticated technology; a detailedgrasp of current and pro jected datacenter capacity requirements; financialacumen; and cooperation of seniorcorporate leadership.

Data centers represent a uniquecombination of financial, IT and realestate resources. While all three must beevaluated, the weight or prioritization ofeach will vary based on your specificneeds. It is essential to identify the optimalblend or “sweet spot” of finance, IT andreal estate. Many projects and programshave failed for not having done so.

opTImIzIng THe place wHere fInance,

IT and facIlITIes converge. Here’s How.

by Mike Duffy

What’s your ideal mix? (Note: allthree areas are in flux and willchange based on evolving businessneeds and new technology.)

60

HOW TO HIT THESWEET SPOT BETWEENFINANCE, IT ANDREAL ESTATEYears of Data CenterDevelopment Experience

Having spent over a decade at Microsoftimmersed in data center design andconstruction, I learned the importance ofad dressing these complex, interrelatedissues to achieve a successful outcomefor all stakeholders. As SeniorDevelopment Manager for Data Centers,I was in charge of creating andmanaging Microsoft’s multi-billion dollarexpansion program. Working on four ofthe ten largest data centers in the world,my responsibilities included financialmanagement, site selec tion, negotiatingtax and utility incentives, and all facets ofdesign, procurement and construction.During my tenure Micro-soft addedseven data centers totaling almost 160Megawatts of critical capacity.

It was a great experience. I worked withreally smart people and gained visibilityinto the emerging data center industry.We defined the best way to build datacenters internally, and we studied thecompetition. What is Google doing?Yahoo? What are their technologies?Locations? We also explored leasedfacilities in the U.S., Europe and Asia;large service firms such as Digital RealtyTrust and DuPont Fabros; andtechnology solutions with industryleaders such as Intel.

I brought all this experience in foundingDuffy Development Services, a datacenter consulting firm in the PacificNorthwest. I’d learned the importance ofa detailed vision of a ‘Dream DataCenter:’ the most important features,options and priorities for engineering,power sourcing, advanced technology,future flexibility, location, security andfinancing. The goal? To align businessand technical needs to the optimal datacenter solution.

When The Benaroya Companyapproached me to consult on theirSeattle area South Hill Data Centerproject, I was a bit skeptical. At MicrosoftI did a lot of site selection, includingpotential repurposing of existingfacilities. The vast majority of propertiesthat owners thought could be convertedto data centers were completelyunworkable. The opposite was true atSouth Hill.

Eliminating Fatal Flaws andTallying Up The Assets

The first task in evaluating a potentialdata center is to look hard for negatives.Any one of three fatal flaws shouldimmedi ately disqualify a site.

1. Not enough reasonably-priced power

2. Inadequate water supply (lessimportant with air-based cooling);

3. Lack of at least one quality fiberprovider (two or more are best, toprovide diversity and pricecompetition).

If a site has no fatal flaws, you begin totally all the positive features. The more,the better. Things to look for include:

[ ] a low total operating cost – notjust the base power rate

[ ] ultra-reliable power – preferablyon-site and easily expandable

[ ] multiple fi ber providers

[ ] scalability and flexibility throughout

[ ] efficient, cost-saving technology

[ ] LEED Gold and/or Energy Starcertification: efficient, sustainable,adaptive, secure, desirable locationwith critical infrastructuremanagement

[ ] one party decision-making, withflexible lease/purchase options

THE BEST DATA CENTERI’VE SEEN IN YEARSI’ve seen a lot of data centers and theone I’m most excited about is the newSouth Hill Data Center south of Seattle.This new data center is a reuse of a$220 million facility, originally built forchip manufacturing with a Class 1 cleanroom. The shell is uniquely designed topush air in a way that a typical datacenter can’t. Intel deployed similar highdensity cooling principles to developtheir Jones Farm facility in Oregon. Theplant at South Hill was clearly anextraordinary property to reposition as adata center.

South Hill Data Center was designedwith operations best practices in mind.Many facilities focus exclusively oncreating capacity, to get it on the marketfast. Operations then become thetenants’ headache.

South Hill offers operational efficienciessuch as loading dock areas for assetmanagement, head-count areas forstaffing, room for IT and facilitiespersonnel, and storage areas so vendorswon’t have to work out of the backs ofvans. It’s a very functional, pleasantenvironment.

In addition, South Hill is the only datacenter I’ve seen that offers a uniquecombination of these three highlydesirable fea tures:


62


WHY DO USERS PLACESUCH A PREMIUM ONTHESE FEATURES?To increase high density capabilities, youmust add additional electrical capacity. Inmost locations this is rarely a problem.It’s adding the necessary coolingcapacity that often becomes a fatal flaw– because it requires financial premiumsand/or significantly impacts efficiency.

South Hill is well-equipped to increasehigh density capabilities, quickly andeconomically. With evaporative coolingand air-side economizers: essentiallyfree cooling capabilities. You don’t payfor mechanical cooling; there’s only theminimal cost of running the fans. Formany data centers, increasing thecapacity of an air-side economizer withan evaporative cooling system ischallenging – sometimes impossible –because of the sheer volume requiredfor air. South Hill can quickly increase

capacity, without disruption and withoutpremium.

South Hill’s in-place scalability isunmatched in other facilities. Initially, thefacility infrastructure is at 150 Watts persquare foot, which can quickly scale upto 260 Watts per square foot. That’s a73% increase in power density, unheardof in data cen ters. South Hill can evenexpand further, without any impact to theserver occupants.

FUTURE-PROOFING THEINVESTMENTScalability is a huge advantage for anoperational data center. Essentially, itfuture-proofs your investment, becausethe de mand for density will continue torise. There are already a number of 80-to-100 Watt facilities that can’t increasetheir power density, because they can’taccommodate a flexible, efficientinfrastructure upgrade. Basically, they arealready outmoded.

Looking into the future, a true “nextgeneration” data center must beequipped to support an industry trendingtoward virtual ization and cloudcomputing.

Virtualization optimizes network and ITresources, getting away from “hands on.”It’s a more managed, holistic, and unifiedapproach that avoids the inefficiencies ofpartially-used servers, partially-usedracks, partially-used data centers.Virtualiza tion optimizes and combinesthose resources, to better utilize ITassets, and to maximize their value andreturn.

With more virtualization, server loads willgo up. And the loads will be morepredictable, more sustainable, and betterman aged. 80 Watt per square foot datacenters will quickly become antiquated.150 Watt data centers will still have aplace, but demand will grow for 200 to250 Watt data centers.

You’ll need a standard high-density podenvironment, with hot-aisle/cold-aislecontainment. An ideal facility is not onlyscal able within its own footprint, but itsinfrastructure and density can also scale.

And with the industry trending towardaddition al servers housed in containers,you’ll need real estate to accommodatethose containers. South Hill offers morethan enough space for containersand/or additional construction, includingan existing cold shell building with plentyof power and fiber.

Apart from engineering, power,technological, safety and securityconsiderations, the bottom line affectingany major finan cial investment is whatwill it cost to run? What will it take tofinance? The power rate at South Hill is5.3 cents and the PUE is 1.32. In atypical facility with a PUE of 2.0, powerrates would need to be 3.5 cents tomatch South Hill.

And South Hill is a turnkey facility, readyfor immediate occupancy. There’s no riskof construction delays and cost-overruns.The easy expandability and scalabilityprotects your investment againstunforeseen financial stress.

CALCULATING TOTALCOST OF OWNERSHIPToo often, Return On Investment (ROI) iscategorized as Capital Cost perMegawatt on a balance sheet. However,companies must also run on a P&Lmodel, with expenses in mind. You mayhave built the cheapest facility in theworld, based on Least Cost PerMegawatt. But if the facility andinfrastructure aren’t efficient, operatingcosts will be high. If you built too muchcapacity – or the wrong density –expenses will be high because utilizationwill be low. And if critical infrastructure –espe cially the mechanical system – isnot efficient, power use will be too high.

ROI should be calculated by Total Costof Ownership rather than just capitalcost, or cost per megawatt. It’s aboutoptimizing processing power and rackspace. How many KW do you actuallyuse relative to what you pay for? If you’reusing just 50% of your resources, evenefficiently, you’re not optimized aroundthe Total Cost of Ownership.

That is why scalability and a low PUE isabsolutely critical. When calculating Total

64

Cost of Ownership, you need to looklong-term. Total Cost of Ownership maybe fine today, but when you need toexpand or scale, it may be compromised.

Looking at the total picture, you’ll likelydetermine that it’s better to go with afacility that has a lower PUE and morescal ability, rather than simply a marginallylower power cost. It’s all about findingthat “sweet spot” where all the factorscon verge. If you overdo any of the criticalfactors to compensate for potentialfuture demands – taking too muchspace, building too high a density or toomuch space, or settling for a poor PUE –you’re stuck. South Hill has an enviablylow PUE and low power costs. It fulfillsall the variables for an optimized TotalCost of Ownership, both short-term andlong-term.

One more thing to consider on thefinancial side – who holds the pursestrings? In my experience, The BenaroyaCompany is a unique owner. They self-finance, so they don’t need to get a bank,

a lender, a partners’ or investors’ input orapproval. The Company is exceptionallypro-active, willing to take calculatedfinancial risk. They know how to movequickly, and can help you do the same.

BACK TO THE SWEET SPOTI hope I’ve given you some usefulguidelines for defining – and ultimatelyfinding – your own data center sweetspot, opti mized for a company’s currentand future financial, IT and real estateneeds.

South Hill Data Center aligns well withtoday’s medium to high densityrequirements. For details, please visitwww.south hilldatacenter.com, or contactDave Vranizan at 425-440-6711.

A CLOSER LOOKSouth Hill Data Center gets the greenlight for anyone seeking an extremelyreliable, secure, highly efficient,

immediately available, state-of-the-artfacility.

Minutes from Sea-Tac Internationalairport, it’s built on solid bedrock, high ona hill, well above the 500 year floodplain. It’s constructed as a building withina building to withstand seismic activity.The site also has a huge water capacityAND multiple fiber providers forredundancy.

Today, South Hill controls 42.5Megawatts of sub-station capacity – thelargest unused block of power in theNorthwest and is expandable to 67.5megawatts. Because the on-site sub-station is dual fed by Alderton and WhiteRiver, the entire region would have tolose power for the sub-station to godown. The world-class grid quality andBonneville power backbone mean thatany short outage — even a milliseconddisruption — would be picked up by UPS.The backup generators will likely neverbe needed.


Mike Duffy is the Principal at Duffy Development Services. He can be reached at [email protected]

66


In a recent report titled “The DigitalUniverse”, industry analyst firm IDCdiscusses the explosion of digitalinformation and how it is causingcompanies to rethink how they design,construct and manage data centers. Aswe live in an electronic world, datacaptured in the form of e-mails,documents, pictures, videos and socialmedia all contribute to this informationexplosion. IDC estimates that, in 2009,this “Digital Universe” came in at awhopping 0.8 zetabytes and that numberis expected to hit 35 zetabytes by 2020.That is enough data to fill a stack of DVDsreaching halfway to Mars.

But what does this have to do with going

green and energy efficiency? As moreinformation requires more IT systems, thismeans more hardware; and morehardware means more energy, which is akiller for green initiatives. Industry analystfirm Pike Research estimates the ITindustry is responsible for approximately2 percent of the world’s carbon emissions– with increasingly complex data centersat the core of this growth. Additionally, arecent report by the EPA, “Report toCongress on Server and Data EnergyEfficiency,” estimates that energyconsumption by servers and data centersdoubled in the past five years and isexpected to do so again over the next five– to more than 100 billion kWh, at a costof $7.4 billion annually.

With more stringent governmentregulations, the need for data centercompanies to reduce energy costs andcarbon emissions is essential. PikeResearch also estimates that investmentsin greener data centers will explode overthe next five years – reaching $41.4 billionby 2015. With a majority of data centerservers operating at only 4% averageutilization (according to the GGT GreenExchange), there is a huge opportunity forcompanies to power forward with energyreduction through better design andoperation, such as IT virtualization and theinstallation of energy efficient electrical/mechanical systems.

The GreenRevolution:

Keeping yourData Centersat MaximumEfficiency

by Jim Embley

Jim Embley is CEO at Rubicon Professional Services. He can be reached at [email protected]


Best Practices: implementing an EnergyManagement Solution

So, how do we get there? It’s clear thatcompanies need to exploreimplementing comprehensive energymanagement strategies and solutions toexploit these opportunities. Thesesolutions are designed to help saveenergy and reduce costs across allphases of the critical facility lifecycle.This service needs to go beyondstandard energy efficiency audits tointegrate a myriad of environmental,technical, legal and administrativeenergy considerations into acomprehensive energy strategy.

To implement successfully, companiesneed to employ both a holistic andflexible approach – one that addressesthe entire facility portfolio whileaccommodating regional differencesand regulatory change. Effective energymanagement initiatives that addressyour various facility portfolios haveproven to reduce Op-ex and Cap-exexpenditures, improve efficiencies andgenerate income. To this end, key facetsto ANY mission critical energymanagement strategy should include:

• Energy Rebates and Incentives:There are literally $100,000,000’s ofincentives that exist for critical facilities.Companies can obtain funding viarebates and cost-share programs fromboth utility and government agencies.In addition, data centers should exploitthese opportunities, as well as evaluate“double-dip” incentive opportunitiesacross multiple funding organizations.

• Demand Response (DR): Incommitting to remove load from gridpower to onsite generation, mostmission critical companies can realize$100,000’s in annual reoccurringrevenue whether or not a DR event iseven called. Because most servicelevel agreements require critical facilityproviders to possess and maintainonsite generation, datacenters areperfect candidates to participate in DR

programs. However, like mostopportunities, the devil is in the detailsand DR is no different. Therefore,companies must possess or retain theexpertise of a partner that:understands regional generatorpermitting requirements; works withfacility operations to ensure powerdistribution systems can effectivelysupport load transfer; and has aworking knowledge of DR providers tostreamline implementation andmaximize revenue.

• Demand Side Management &Sustainability Initiatives: Missioncritical companies must first establisha baseline by conducting effectiveenergy audits and efficiencyassessments that quantify and suggeststrategies to reduce energy spend,which is one of the largest budget lineitems. This baseline will then help guidea data center to effectively manageequipment specifications andinstallations as well as create payback,total cost of ownership, and powerusage effectiveness (PUE)calculations that gauge its progress.

• Reporting & Ongoing Management:Strategy is pointless unless successcan be quantified and measured.Therefore, energy managementprograms need dedicated resources tomonitor, manage and regularly updatethe various initiatives across a facilityportfolio. This is key to not onlyachieving established goals but alsoproving to senior management theeffectiveness of a company’s energymanagement plan.

With the digital universe exploding,comprehensive energy managementstrategies must be implemented to staycompetitive. The world is changing andyou need to change with it. Significantefficiency and revenue opportunities arethere for the companies that driveenergy management initiatives. Startwith a plan, retain the expertise toimplement the plan, and stick with itbecause the results will give yourorganization a competitive advantage aswell as cash to the bottom line.

There are $100,000,000’s of energyincentives throughout the US for missioncritical data centers, including:

• NYSERDA – New York State Energy andResearch Development Authority

• Con Ed (NYC Utility)• NJCEP – New Jersey’s Clean EnergyProgram

• NStar (Boston Utility)• Com Ed (Chicago Utility)• Oncor (Dallas Utility)• Centerpoint (Houston Utility)• Seattle City Light (Seattle Utility)• Southern California Edison (Utility)

Additionally, a data center administrator canfind many incentive programs through theDatabase of State Incentives for Renewable& Efficiency (DSIRE) http://dsireusa.org/.Established in 1995, the Database of StateIncentives for Renewable & Efficiency(DSIRE) is an ongoing project of the NorthCarolina Solar Center and the InterstateRenewable Energy Council (IREC). It isfunded by the U.S. Department of Energy’sOffice of Energy Efficiency and RenewableEnergy (EERE), primarily through the Officeof Planning, Budget and Analysis (PBA). Thesite is administered by the NationalRenewable Energy Laboratory (NREL), whichis operated for DOE by the Alliance forSustainable Energy, LLC.

Identifying the programs is easy. The realquestion is how to effectively obtain cashfunding from these organizations in a timelymanner while meeting both the goals of theapplicant and the agency administering theprogram.

Recommended Best Practices:

• The incentive agencies must achieve kWreduction goals and issue amounts offunding that are proportionate to obtainingsuch goals. To this end, many incentiveagencies undergo annual audits by thestate’s attorney general’s office or publicutility commission/service.

• Although similar in objective, each incentivebody throughout the county has differentinitiatives, processes, timelines, and amountsof funding. For example, there are incentivesfor energy site assessments, energy efficientequipment installs, and demand responseequipment installs. All require varyingdegrees of effort and documentation onbehalf of the applicant and have varyingfunding levels/caps.

Energy incentives at Your Fingertips

• Applicants must be mindful that funding levels for each agency changefrom year-to-year. Rubicon informs clients of funding levels prior toinvesting time in creating and submitting applications.

• To obtain funding, organizations must do more than simply fill out a form.Partner with a vendor that can advise on timelines, stages, gates, andmeasurements/verifications required to obtain funding and incorporatethese in the overall facility infrastructure plan. For example, most agencieswill disqualify funding if energy efficient equipment has been ordered priorto submitting the incentive application!! To prevent disqualifications, only adedicated resource with knowledge and experience can manage multipleincentive agencies and the various funding initiatives.

• In addition, agencies subcontract the administration of most incentiveprograms to third parties because they typically do not have the in-housepersonnel to manage these initiatives 100%.

Therefore data centers need to retain resources (either through in-housetalent or outside partners) with expertise and experience in dealing not onlywith the incentive agencies, but also their subcontractors, and have anunderstanding of the criteria and amounts of funding for the variousinitiatives that each incentive agency offers. To acquire this level of expertise– especially for mission critical companies with a national footprint acrossvarious incentive bodies — data centers need a resource dedicated toidentifying incentive opportunities, knowing what initiatives bring in themaximum amount of funding, and shepherding the process from creatingthe application to cashing the check.

Differentiating legitimate programs from a potential scam

• As a rule of thumb, all energy incentive programs are either administeredby a local utility (e.g., Con Ed) or a state agency (e.g., NYSERDA).

• To validate the incentive program, one should at least contact theapplicable state’s Public Service Commission/Service or the applicablestate’s Attorney General’s Office, as well as reference http://dsireusa.org/.

Examples of “double-dip” incentive opportunities

• In New York, a data center can receive capital funding for demandresponse equipment installs from NYSERDA, while receiving demandresponse quarterly payments from NYISO.

• In several states, a data center could receive funding from two differentagencies for the same site. For example, can include a new constructionincentive from the local utility concerning the build out of a suite, and anenergy efficiency incentive from a state agency for upgrading equipmentin another part of their site.

73


INTERESTED IN PRESENTINGAT THE FALL CONFERENCE?Visit www.7x24exchange.org and download the Call for Presentations

DEADLINE: AUGUST 5TH

SUBMIT AN ARTICLE FOR THE FALL 7x24MAGAZINEVisit www.7x24exchange.org and download the Call for Articles

DEADLINE: JULY 8TH

2 0 1 1 S P R I N GEnd-to-End rEliability

MISSION CRITICAL FACILITIES:

emergingtrends

FALL 2011November 13 – 16, 2011Arizona BiltmorePhoenix, AZ

SPRING 2012June 10 – 13, 2012Hilton Orlando Bonnet CreekOrlando, FL

Interested in presenting at 7x24 Exchange? Complete the Call for Presentations at www.7x24exchange.org or call 646-486-3818 x104

S A V E T H E D AT E ! ! !

For information about sponsoring a 7x24 Exchange eventplease contact Brandon Dolci, CMP at (646) 486-3818 x108

June 12–15, 2011 Hilton Orlando Bonnet Creek Orlando, FL

74

75

76

2011 SPRINGCONFERENCEHIGHLIGHTS

For the complete Spring Conference program and registration information please visit7x24exchange.org or call (646) 486-3818.

The Spring Conference themed “End-to-End Reliability: Emerging Trends” will be heldJune 12-15 at the Hilton Bonnet Creek in Orlando, FL. The Spring Conference willfeature a Data Center Energy Practitioner workshop and exam, compelling keynotes,concurrent sessions, sessions on solar flares, modular data centers, emerging trends, a spectacular vendor event, and more...

Robert F. Kennedy, Jr., Visionary, Environmental Business Leader &Advocate, will kick off the conference with a session entitled “Green GoldRush: A Vision for Energy Independence, Jobs and National Wealth”.

Yahoo’s Vice President of Operations, Scott Noteboom, will open thesecond day with a keynote presentation entitled “Right Place. Right Way.Right Cost: Evolving Efficiency in Data Factory Design and Operations”.

Bob Cashner, SVP of Corporate Properties Group for Wells Fargo willprovide the closing keynote “End User’s Perspective on SuccessfullyNavigating Data Center Emerging Trends”.

In keeping with the theme, additional presentations on emerging trends will bedelivered with topics such as:

• eBay – Data Center Goes Gold

• NASA – The Influence of Solar Flares and Solar Storms: Why We Should Care About Space Weather

• Green Grid – Economic and Environmental Value in Business Computing

• Uptime – Data Center Issues: Past, Present & Phuture

• United Parcel Service – Open Heart Surgery on the World’s First Tier IV Data Center

• University of Chicago – What’s Old is New Again: Integrating a High Density DataCenter into a Historic Building

In addition to enhanced programming 7x24 Exchange International presents “AnEvening At Sea World”. Join us on as we venture off property to Sea World whereattendees will be treated to cocktails, dinner, musical entertainment and a special visitfrom live animals followed by the opportunity to enjoy the attractions at Sea World.

This event has been made possible thanks to the following partners:

ABB, Active Power, ASCO, Balfour Beatty, Chil-Pak, Caterpillar, ComRent,

Cummins Power Generation, Cyberex, Data Aire Inc., DPR Construction,

Eaton, Emerson Network Power, FieldView, GE, Gilbane, idGroup, Kohler,

Layer Zero, Mitsubishi, MTU, Opengate, PDI, Russelectric, S&C,

Schneider, Siemens, Skanska Mission Critical, Staco Energy, Starline,

Stulz, Syska Hennessy, Truland, United Meta Products, Virginia Economic

Development Partnership, Walker Engineering, Whiting-Turner

77

2011 SPRINGCONFERENCE

CORPORATELEADERSHIP

PROGRAMPARTNERS(AT PRESS TIME)

MEDIAPARTNERS

GOLDPARTNERS

PLATINUMPARTNERS

MITSUBISHI ELECTRIC

UPS Division

SILVERPARTNERS

BRONZEPARTNERS

79

WWW.7X24EXCHANGE.ORG

D E L A W A R E V A L L E Y C H A P T E R D E L A W A R E V A L L E Y C H A P T E R D E L A W A R E V A L L EY C H A P T E R D E L A W A R E V A L L E Y C H A P T E R D E L A W A R E V A L L E Y C H A P T E R D E L A WA R E V A L L E Y C H A P T E R D E L A W A R E V A L L E Y C H A P T E R D E L A W A R E V A L L E Y C H A PT E R D E L A W A R E V A L L E Y C H A P T E R D E L A W A R E V A L L E Y C H A P T E R D E L A W A R E V AL L E Y C H A P T E R D E L A W A R E V A L L E Y C H A P T E R D E L A W A R E V A L L E Y C H A P T E R D EL A W A R E V A L L E Y C H A P T E R D E L A W A R E V A L L E Y C H A P T E R D E L A W A R E V A L L E Y CH A P T E R D E L A W A R E V A L L E Y C H A P T E R D E L A W A R E V A L L E Y C H A P T E R D E L A W A RE V A L L E Y C H A P T E R D E L A W A R E V A L L E Y C H A P T E R D E L A W A R E V A L L E Y C H A P T ER D E L A W A R E V A L L E Y C H A P T E R D E L A W A R E V A L L E Y C H A P T E R D E L A W A R E V A L LE Y C H A P T E R D E L A W A R E V A L L E Y C H A P T E R D E L A W A R E V A L L E Y C H A P T E R D E L AW A R E V A L L E Y C H A P T E R D E L A W A R E V A L L E Y C H A P T E R D E L A W A R E V A L L E Y C H AP T E R D E L A W A R E V A L L E Y C H A P T E R D E L A W A R E V A L L E Y C H A P T E R D E L A W A R E VA L L E Y C H A P T E R D E L A W A R E V A L L E Y C H A P T E R D E L A W A R E V A L L E Y C H A P T E R DE L A W A R E V A L L E Y C H A P T E R D E L A W A R E V A L L E Y C H A P T E R D E L A W A R E V A L L E YC H A P T E R D E L A W A R E V A L L E Y C H A P T E R D E L A W A R E V A L L E Y C H A P T E R D E L A W AR E V A L L E Y C H A P T E R D E L A W A R E V A L L E Y C H A P T E R D E L A W A R E V A L L E Y C H A P TE R D E L A W A R E V A L L E Y C H A P T E R D E L A W A R E V A L L E Y C H A P T E R D E L A W A R E V A LL E Y C H A P T E R D E L A W A R E V A L L E Y C H A P T E R D E L A W A R E V A L L E Y C H A P T E R D E LA W A R E V A L L E Y C H A P T E R D E L A W A R E V A L L E Y C H A P T E R D E L A W A R E V A L L E Y C HA P T E R D E L A W A R E V A L L E Y C H A P T E R D E L A W A R E V A L L E Y C H A P T E R D E L A W A R EV A L L E Y C H A P T E R D E L A W A R E V A L L E Y C H A P T E R D E L A W A R E V A L L E Y C H A P T E RD E L A W A R E V A L L E Y C H A P T E R

Delaware Valley Chapter

1st Row: (L to R) Gary Aron, Donna Aron, Tom Reusche, Erin Reusche, Donna Manley, Mark Jacobs

2nd Row: (L to R) Robin Rinaldi, Diane Werner, Louise Marvel, Ruth Michel, Phil Michel

3rd Row: (L to R) Mike Rinaldi, Gary Hill, Rich Werner, Bill Leedecke, Frank Reidy, Doris Leedecke

1st Row: (L to R) Gary Aron, Donna Manley, Frank Reidy, Phil Michel

2nd Row: (L to R) Tom Reusche, Rich Werner

3rd Row: (L to R) Mike Rinaldi, Gary Hill, Bill Leedecke

The 7x24 Exchange Delaware Valley Chapter celebrated its 50th Chapter meeting at the historic Union League ofPhiladelphia. The event featured two presentations by John Diamond of DAS Associates, and a former Board Member ofthe Delaware Valley Chapter, “7x24 Exchange Delaware Valley Chapter at 50: A Look Back, A Look Ahead,” and “DataCenters 2011: State-Of-The-Art, From Micro To Mega.” 130 people attended the black tie optional event, whichfeatured a jazz quartet from University of the Arts in Philadelphia. Frank Reidy was honored for his service as PastPresident of The Board of Directors.

CHAPTER NEWS


W A S H I N G T O N D C C H A P T E R W A S H I N G T O N D C C H A P T E R W A S H I N G T O N D C C H AP T E R W A S H I N G T O N D C C H A P T E R W A S H I N G T O N D C C H A P T E R W A S H I N G T O N D CC H A P T E R W A S H I N G T O N D C C H A P T E R W A S H I N G T O N D C C H A P T E R W A S H I N G T ON D C C H A P T E R W A S H I N G T O N D C C H A P T E R W A S H I N G T O N D C C H A P T E R W A S H I NG T O N D C C H A P T E R W A S H I N G T O N D C C H A P T E R W A S H I N G T O N D C C H A P T E R W AS H I N G T O N D C C H A P T E R W A S H I N G T O N D C C H A P T E R W A S H I N G T O N D C C H A P T ER W A S H I N G T O N D C C H A P T E R W A S H I N G T O N D C C H A P T E R W A S H I N G T O N D C C HA P T E R W A S H I N G T O N D C C H A P T E R W A S H I N G T O N D C C H A P T E R W A S H I N G T O N DC C H A P T E R W A S H I N G T O N D C C H A P T E R W A S H I N G T O N D C C H A P T E R W A S H I N G TO N D C C H A P T E R W A S H I N G T O N D C C H A P T E R W A S H I N G T O N D C C H A P T E R W A S HI N G T O N D C C H A P T E R W A S H I N G T O N D C C H A P T E R W A S H I N G T O N D C C H A P T E R WA S H I N G T O N D C C H A P T E R W A S H I N G T O N D C C H A P T E R W A S H I N G T O N D C C H A P TE R W A S H I N G T O N D C C H A P T E R W A S H I N G T O N D C C H A P T E R W A S H I N G T O N D C C HA P T E R W A S H I N G T O N D C C H A P T E R W A S H I N G T O N D C C H A P T E R W A S H I N G T O N DC C H A P T E R W A S H I N G T O N D C C H A P T E R W A S H I N G T O N D C C H A P T E R W A S H I N G TO N D C C H A P T E R W A S H I N G T O N D C C H A P T E R W A S H I N G T O N D C C H A P T E R W A S HI N G T O N D C C H A P T E R W A S H I N G T O N D C C H A P T E R W A S H I N G T O N D C C H A P T E R WA S H I N G T O N D C C H A P T E R W A S H I N G T O N D C C H A P T E R W A S H I N G T O N D C C H A P TE R W A S H I N G T O N D C C H A P T E R W A S H I N G T O N D C C H A P T E R W A S H I N G T O N D C C HA P T E R W A S H I N G T O N D C C H A P T E R W A S H I N G T O N D C C H A P T E R W A S H I N G T O N DC C H A P T E R

Washington DC Chapter

80

Celebrated its one year anniversary withover 100 members and guests at our 1stAnnual Membership Drive and VendorNight on March 21st. Thanks to our vendortable sponsors who made this a hugesuccess: Active Power, BTech, ComRent,Digital Realty Trust, HorizonLineTechnologies, Power Protection Unlimited,RTKL, Server Technology, T5, UGLServices and Whiting-Turner.

On April 8th, seventy-five of our chapter membersand guests were given a tour of Fannie Mae’s Tier4, LEED Certified Data Center from theprofessionals that designed, built and are currentlymaintaining it. An open panel discussion was heldat the tour conclusion allowing members to hearthe current and future sate of data center design.Due to the huge success of this event, we’ll betouring Dupont Fabros’ facility in Ashburn for ourJune event. (info available at www.7x24dc.org)

From March 21st Membership Drive: Steve Mantua & Chris Maruca both with CTS Services,John Fritz of HITT Contracting, Melanie Naugle EDG2, Ben O’Conner CTS Services andTommy Inscoe with Power Solutions.

Membership Drive: Jonathan Litvany – Digital Realty Trustand Chuck Meyer – TM Technology Partners

CHAPTER NEWS

BECOME INVOLVED IN YOUR LOCAL 7x24 CHAPTER

BCAB

SK

ND

ON

MT

WYSD

NE

CO

NM

KS

LA

OK

WI

IA

IL

TN

MS

TX

AR

MO

MN

ME

ID

WA

OR

CA

NV

AZ

UT

MI

NCKY

INOH

SCGA

FL

AL

WV

MARICTNJDEMDDC

NHVT

NY

PA

VA

Arizona

Atlanta

The Carolinas

Delaware Valley

Florida

Greater Washington DC Area

Lake Michigan Region

Lone Star (Dallas)

Metro New York

Midwest (Kansas City, MO)

New England Area

Northern California

Northwest (Seattle, WA)

Rocky Mountain

Southeast Michigan

Southern California

Texas South

VISIT WWW.7x24EXCHANGE.ORG TODAY TO PARTICIPATE IN YOUR LOCAL CHAPTER

Attention end users and vendors...

INTERNATIONAL CHAPTERS INCLUDE:

MEMBER ADVERTIS

ING RATE CARD

7x24MAGAZINE OPPORTUNITIESFollowing are the Editorial Guidelines together with the Member Advertising Rate Card. Advertisersinterested in placing an ad may fax the insertion order to 7x24 Exchange at 212.645.1147 or emailto [email protected]. Questions? Please call Jeremy O’Rourke at 646.486.3818x109.

Advertiser indemnifies 7x24 Exchange against losses or liabilities arising from this advertising. 7x24 Exchange assumes no liability whatsoever,except to the extent of a one time paid advertisement of the same specification, in the next or similar publication, if any proven or admittederrors or omissions have occurred. Payment is due upon receipt of the invoice. Interest shall be charged at 2% per month compounded to yield26.82% per year on overdue accounts. Revisions to previously submitted ad copy are subject to additional charges. A charge of $30.00 will belevied for returned checks. In the event of a contract cancellation, the advertiser or agency agrees to repay 7x24 Exchange any discountsgranted for multiple insertions less any discount applicable for the number of insertions completed in the contract. All cancellations must bereceived in writing prior to the advertising sales deadline. All premium positions are non-cancelable. Prices are net of agency commission.

BLACK AND WHITE RATES

Size 1X 2X 3X

Full Page $1,500 $1,300 $1,1002/3 Page 1,100 1,000 9001/2 Page Island 900 800 7001/2 Page 700 600 5501/3 Page 600 550 5001/4 Page 500 450 400

COVERS & PREMIUM POSITIONS – INCLUDES 4 COLOR

Size 1X 2X 3X

DPS $5,000 $4,500 $4,0002nd / 3rd Cover 2,500 2,200 2,0004th Cover 3,500 2,750 2,500

Full

Page

1/4

Verti

cal

1/4

Hor

izon

tal

1/3

Verti

cal

1/3

Hor

izon

tal

1/2

Verti

cal

1/2

Hor

izon

tal

1/2

Islan

d

2/3

Verti

cal

2/3

Hor

izon

tal

Live Area: 7.5” x 10”Trim Size: 8.5” x 11”Bleed Size: 8.75” x 11.25”Halftone Screen: 133 lines up to 150 linesDPS Mechanical Requirements:Live Area: 16” x 10”Trim Size: 17” x 11”Bleed Size: 17.25” x 11.25”Halftone Screen: 133 lines up to 150 lines

8 1/2” x 11” MECHANICAL REQUIREMENTS

COLOR RATESProcess Color (4/c): $900

Revisions and Proofs: $50

Position Guarantee: 15% premium

*Non-Members add 40% to all rates

NON-BLEED AD DIMENSIONSSize Width LengthFull Page 7.5” 10”2/3 Horizontal 7.5” 6.5”2/3 Vertical 5” 10”1/2 Island 4.875” 7.5”1/2 Horizontal 7.5” 4.875”1/2 Vertical 3.625” 10”1/3 Horizontal 7.5” 3.25”1/3 Vertical 2.5” 10”1/4 Horizontal 4.5” 3.25”1/4 Vertical 3.25” 4.5”

EDITORIAL GUIDELINESManuscript specifications: Feature articles vary in length from 500 to 2,000words. While we accept articles in a variety of formats, it prefers to receivematerials on CD. All articles must be received by the deadline to beconsidered for a specific issue. Material submitted after the deadline will beconsidered for the following issue.

Bylines: All articles should include a brief (1-2 sentence) author biographicalsketch at the end of the article, that includes the author’s name, title,affiliation, address, and phone number. Photos of authors are never used. Wedo not pay authors for contributions.

Visuals: Authors are encouraged to submit photographs and charts, graphs,or other illustration that will help readers understand the process beingdescribed, though it does not guarantee that visuals will be used with thearticle. Submit all charts, graphs, and other artwork separately; do notincorporate them in the body of the article. Indicate caption materialseparately. We reserve the right to publish submitted visuals.

Editorial proceduresAll articles are reviewed for suitability. Accepted materials are then edited forgrammar and to conform with our editorial style. All attempts are made topreserve the author’s writing style, however, we have the right to edit for style,clarity, and to fit space allotments, and to make final selection on headlines,subheads, and graphic treatment. Manuscript submission implies authoragreement with 7x24 Exchange’s Editorial Policies.

CopyrightWe require first serial rights for submitted articles. This means the author(s)grant us the right to publish the article for the first time. We also requestpermission for electronic distribution on 7x24 Exchange’s web site,www.7x24exchange.org.

DisclaimerThe responsibility for accuracy remains with the author. The opinions andinformation in bylined articles in this publication are those of the authors anddo not necessarily reflect those of the Officers and Board of Directors of 7x24Exchange.

Press optimized PDF or EPS accepted.No JPEG or TIFF accepted.

Convert all spot colors to CMYK.

Do not use 4 color black. All 4 colorblack will be replaced with 100%black at the advertiser’s expense.

Fonts must be embedded or outlined.

Maximum ink density 280%

83

Documents

Space Weather: A Brief Overview - 7x24 Exchange …Syska Hennessy Group Administrative Director Kathleen A. Dolci 646.486.3818 x103 Membership & Education Tara Oehlmann, Ed.M. 646.486.3818