Estimating Power Consumption In Indian Data Centers

By Author: Greenfield Software
Total Articles: 5
Comment this article

Introduction
Data Centers around the world have become major power consumers. It has been estimated in the United States and European Union (see Table 1) thatData Centers consume over 2% of utility-produced electrical power and the consumption is growing
at an alarming rate of over 15% per annum! Over and above this, It have
100% back-up power mostly operated on diesel.

Table 1: Data on Power Consumption and Production(figures in Billion kwh)
Region 2007 2011(e) Utility Produced Power (2010)
United States 61 100 4120
European Union 56 104 3100
Sources: US EPA and European Commission JRC on Data Center Power Consumption
US Census Bureau 2012 Statistical Abstract; Eurostat: Statistics Explained on Utility Power Produced

The good news is that since 2007, when the problem was first measured and highlighted, It in North America and Europe began to take measures to control this growth through more efficient HVAC systems, higher energy efficient servers, renewable energy options, better layouts and implementing Data Center Infrastructure Management software ...
... DCIM software helps to track device-wise power usage and makes recommendations for improving total energy efficiency across the Data Center.
In India, on the other hand, we do not have such baseline data of power consumption in Data Centers. This paper aims to make a case why it has become imperative to do so from a policy making perspective. Base liningwill help to set benchmarks and evolve necessary guidelines for energy efficiency in this emerging but power-intensive sector by different parameters.

Emergence of Indian Data Centers
Growth: With nearly 8% per annumaverage economic growth (until last fiscal year), there was an inevitability of growth in data center capacity of existing facilities as well due to higher transactional volumes and digital data growth. Compounded with this were three other factors: (i) growth in e-commerce; (ii) growth in e-Governance; and (iii) growth in outsourcing to India for computing facilities.
Topological Change: Businesses and Government cannot afford any risks to their computing infrastructure. Hence nearly every organization that has a primary Data Center (DC) also has a Disaster Recovery (DR) DC. Also Tier 3 and Tier 4 DCs (irrespective of primary or DR) are built with extreme redundancies to meet 99.99% uptime requirements.Logically, for every unit of computing work done, there is a 4x multiplier effect in asset base and >2x in power consumption. It is to be noted here that non-computing assets (viz. DC assets excluding servers, storage and networks) account for nearly 50% of total power consumption in most software, with primary load coming from cooling requirements.
Energy Savings Potential in Indian Data Centers
An accepted formula for measuring energy efficiency in a Data Center is Power Usage Effectiveness (PUE) which is defined as:
Total Power Consumption in the Data Center
Total Power Consumption by IT Equipment

The ideal state is for this ratio is 1.0. The most efficient Data software in the world have achieved between 1.15 – 1.20. The BEE-CII Manual on Data Center Energy Efficiency states “Typically, for a conventional datacenter with operating Power Usage Effectiveness (PUE) of 2.0, theminimum energy saving potential by adoption of latest technologies is between 25 to 30%”. Of the savings, while nearly 60% can come from server-related efficiencies like virtualization, there’s another 7% that may be derived from more efficient cooling technologies.

Why and What of Estimation?
Energy usage is crudely monitored, if at all. Rising power consumption compounded with escalating energy prices, makes it important to measure and manage power use through Building Management Systems (BMS) and DCIM software.
• Measure power consumption and determine key metrics like Power Usage Effectiveness (PUE), Watts per Rack,and Watts per square foot.
• Analyze power consumption patterns with respect to different segments, size and redundancy levels.
• Formulate benchmarks for Indian Data Centers per above mentioned parameters and compare with international standards.
• Estimate power consumption in for 2012, which can then be treated as baseline year.
• Estimate annual growth and forecast power consumption over next 4 years.

Benefits
The study will provide energy policy makers critical data to:
a) Formulate audit guidelines to certify on energy efficiencyusing PUE and other measures like watt/square foot and watt per rack
b) Provide incentives or penalties based on actuals versus benchmarks
c) Recommend ways to reduce power consumption in the data center through policy interventions
d) Arrest avoidable growth in power consumption, reduce pressure on power systems and harness the savings towards other usage
e) Formulate incentive policy for installing back-up power units based on non-conventional and/or renewable energy and injecting surplus generation to the grid.
For more information please visit http://www.greenfieldsoft.com