Mak­ing Data Cen­ters En­ergy Ef­fi­cient

The data cen­ter is an ecosys­tem. A ju­di­cious blend of tech­nolo­gies blends to­gether to cre­ate a suc­cess­ful one

Dataquest - - C9ONTENTS -

Ac­cord­ing to U.S. Department of En­ergy’s Of­fice of En­ergy Ef­fi­ciency and Re­new­able En­ergy (EERE), it says, “Data cen­ters con­sume large amounts of en­ergy to run and main­tain their com­puter sys­tems, servers, and as­so­ci­ated high­per­for­mance com­po­nents—up to 3% of all U.S. elec­tric­ity pow­ers data cen­ters. Data cen­ters can be­come more en­ergy ef­fi­cient by in­cor­po­rat­ing fea­tures like pow­er­sav­ing “stand-by” modes, en­ergy mon­i­tor­ing soft­ware, and ef­fi­cient cool­ing sys­tems. These ef­fi­ciency im­prove­ments can pro­duce sig­nif­i­cant en­ergy sav­ings, re­duce the load on the electric grid, and help pro­tect the na­tion.

PRI­OR­I­TIZ­ING EN­ERGY EF­FI­CIENCY

In­dus­try ex­perts do agree. Ravi Raj, Brand Head, Di­rec­tor Sales & Sup­port at NetRack En­clo­sures says, “As data cen­ter is the back­bone of any or­ga­ni­za­tion and it re­quires a lot of at­ten­tion while choos­ing the right set of com­po­nents to build a cost-ef­fec­tive and fu­ture-proof in­fra­struc­ture. Ac­cord­ing to Gart­ner, ap­prox­i­mately 10% of data cen­ter op­er­at­ing ex­pen­di­ture (OPEX) is power, and power is likely to be about 15% of data cen­ter OPEX within five years.”

Clearly, the data cen­ter is an ecosys­tem. A ju­di­cious blend of tech­nolo­gies blends to­gether to cre­ate a suc- cess­ful one. Re­flect­ing on that M Muthuku­mar, VP, En­gi­neer­ing, In­dia & SAARC, Ju­niper Net­works, IEC says, “Op­ti­mal and ef­fi­cient utiliza­tion of the com­pute and stor­age re­sources and net­work in­fra­struc­ture are key to en­ergy ef­fi­ciency. In­stead of ded­i­cated and over-pro­vi­sion­ing of re­sources a shared un­der­lay in­fra­struc­ture with vir­tu­al­iza­tion tech­nol­ogy en­ables the most value per unit of in­vest­ment. Net­work in­fra­struc­ture ven­dors need to drive green tech­nolo­gies to drive “watt per bit” as a key pa­ram­e­ter in their ar­chi­tec­ture and de­sign.”

Ex­perts agree that de­sign, en­gi­neer­ing, and plan­ning play very im­por­tant roles as data cen­ter in­fra­struc­ture needs to be more gran­u­lar and mod­u­lar. This starts from the site selec­tion process as cli­mate zone of the lo­ca­tion plays an im­por­tant role in en­sur­ing en­ergy ef­fi­ciency. The right plan­ning not only de­fers cap­i­tal ex­pen­di­ture but also im­proves en­ergy ef­fi­ciency by higher space and power utiliza­tion. Hence, white space ca­pac­ity plan­ning be­comes ex­tremely crit­i­cal to main­tain the right bal­ance and avoid over-pro­vi­sion­ing.

More­over one needs to drill down to Elec­tro-Me­chan­i­cal (MEP) sys­tem level and com­po­nent level ef­fi­ciency pa­ram­e­ters. Putting this thought in the back­drop, Prasanna Saram­bale, CEO – Data Cen­ter Busi­ness and Head – Group BD, Ster­ling and Wil­son says, “De­pend­ing on

the IT load per rack, Cool­ing in data cen­ters ac­counts for ma­jor over­heads in terms of en­ergy con­sump­tion mak­ing it the prime focus dur­ing En­gi­neer­ing, Pro­cure­ment and Con­struc­tion (EPC) of the phys­i­cal in­fra­struc­ture. Amer­i­can So­ci­ety of Heat­ing, Re­frig­er­a­tion, and AirCon­di­tion­ing En­gi­neers (ASHRAE) TC 9.9 ((the tech­ni­cal com­mit­tee fo­cused on data cen­ters) pro­vides a good in­sight in terms of en­vi­ron­men­tal con­di­tions that need to be main­tained in a data cen­ter area. ASHRAE has also brought in­dus­try bench­marks in terms of en­ergy per­for­mance for phys­i­cal in­fra­struc­tures, en­sur­ing that data cen­ters com­ply with a cer­tain stan­dard with re­spect to their heat load en­ergy con­sump­tion.”

“Fi­nally, one of the most im­por­tant el­e­ments - the Power Utiliza­tion ef­fi­ciency (PUE) of a data cen­ter fa­cil­ity, de­pends on how the en­tire plant is op­er­ated and main­tained on a 24x7 ba­sis. Con­sid­er­ing these crit­i­cal as­pects, it is im­per­a­tive to part­ner with a data cen­ter EPC com­pany that can de­sign, plan, im­ple­ment and op­er­ate the data cen­ter ef­fi­ciently through­out its life cy­cle,” adds Saram­bale.

IT’S ALL IN THE DE­SIGN

At the end of the day, the cre­ation of an en­ergy ef­fi­cient data cen­ter lies in a suc­cess­ful de­sign that takes into con- sider­a­tion whole lot of as­pects needed for suc­cess­ful de­liv­ery of com­put­ing re­sources. As Muthuku­mar sums up, “Mas­sive scale, the sim­plic­ity of pro­vi­sion­ing and op­er­a­tion are the fun­da­men­tal de­sign prin­ci­ples of data cen­ters, while re­siliency, se­cu­rity, and elas­tic ca­pac­ity guide the de­sign from a ser­vice of­fer­ing point of view. All of this leads to the tenet of a data cen­ter is “one large com­puter” while there are phys­i­cally dis­crete re­sources - CPU, mem­ory, and Net­work. Or­ches­trat­ing dy­namic work­loads across phys­i­cal re­sources and steer­ing data via the in­ter­con­nect­ing net­work in the most op­ti­mal fash­ion is where SDN tech­nolo­gies play a key role. There are sev­eral ex­am­ples where an en­tire en­ter­prise data cen­ter can be boot­strapped over a pub­lic cloud in a mat­ter of hours as against sev­eral days and months con­sist­ing of de­sign, pro­cure­ment, in­stal­la­tion, and Go-Live.”

Newspapers in English

Newspapers from India

© PressReader. All rights reserved.