You are here

Greening Big Data

With the advent of the Internet of Things, Big Data is driving the large businesses in the world and the penetration of data centers in our lives is set to increase manifold. India’s IT sector contributed about 8% to the GDP in 20141  and is a large consumer of energy (even though official estimates are not available). Sustainability Outlook delves into the operations of a data center to determine how data centers can become greener. 
It is expected that propelled by Modi’s “Digital India” campaign, Data Centers space will grow at a CAGR of over 22% between 2015 and 2020. Indian data centers are projected to occupy 6.3 million sq. ft. by 2017.2 With an average power consumption of 1127 kW per 5000 sq. ft.,3,4 annual energy consumption of data centers will amount to 12,200 TWh by 2017. Studies show that by incorporating best practices, it is possible to reduce data center energy consumption by 40-60%. A 50% reduction in energy consumption would mean annual energy savings of 6,100 TWh, which is equivalent to the energy consumption of 5.6 million Indian households.  5

The need for “Greener” Data centers

Customer Demand
Today, corporates, both large and small are moving to the cloud. For many corporate clients, “Green” becomes an effective differentiating factor while choosing which company to work with. Many organizations buying data center services have become more aware of green operations and demand that their data centers follow these practices as well.  Beroz Gazdar, CSO of Mahindra Group, confirmed the same in a conversation with Sustainabiltiy Outlook earlier this year, “For B2B businesses like Tech Mahindra, we’ve seen that sustainability is a very important factor for clients.” 
Energy Efficiency and Conservation leads to Savings
The business case for energy efficiency in Data Centers has long been proven. Data Centers in India and all over the world have seen significant financial savings with payback periods ranging from a few months to a few years for investment in Energy Efficiency or Energy Conservation measures.
Measures to improve data center sustainability would relate to not just the operational energy and water consumption, but also the building design, resources used in building and maintenance of the data center, waste generation, management and recycling.  But first, to accurately measure, monitor and track progress on these fronts, it is necessary to have the adequate metrics for measurement.
Green Data Centers- Metrics for monitoring
The most common metric used for measuring energy utilization and efficiency in a data center is Power Usage Effectiveness (PUE). Other common metrics used are Green Energy Coefficient (GEC), Material Recycling Ratio (MRR), IT Equipment Energy Efficiency (ITEE), and IT Equipment Utilization (ITUE).

How can Data Centers reduce their Environmental Impacts?

Outlined below are some of the measures a Data Center may take to reduce its environmental impact. The measures have been divided into three stages- Beginner, Improver and Achiever, charting the steps data centers can take in the short, mid and long term.

Measures to improve ITEE and ITUE- Reduce Energy use in IT Equipment 

The following table summarizes some of the ways in which Energy efficiency of IT equipment can be improved:
A study undertaken by the Bureau of Energy Efficiency found out the potential energy savings from various initiatives in data centers in India. The findings are indicated in Figure 1 below.
Figure 1: Percentage of Energy Savings that can be achieved by implementing various initiatives in Data Centers: Source: BEE, CII, Lawrence Berkeley National Labroatory Report of Data Center Efficiency in India 6

Measures to improve PUE - Reduce energy use for HVAC

A study by Lawrence Berkeley National Laboratory showed that the HVAC energy consumption in Indian data centers varied from about 28% to 50% of total data center energy consumption.7 The same study also found that the PUE for Indian data centers varied from 1.7 to 2.1 which indicates that there is a large room for improvement for most data centers.
Resetting the Thermostat

One of the simplest ways to save energy in a data center is to raise the temperature. The American Society of Heating, Refrigerating and Air Conditioning Engineers (ASHRAE) recommends cold aisle temperatures of up to 27°C, which has been found to have no detrimental effect on equipment. Most IT equipment manufacturers spec machines at 32°C or higher, so there is sufficient margin to operate, even in the case of a power failure. 
Manage airflow- Hot and Cold Aisle Containment
Cooling is the largest contributor to facility overhead energy. The most important step in optimizing air flow is preventing hot and cold air from mixing.  In a typical data center, the IT equipment is organized into rows; usually with a “cold aisle” in front - here the cold air enters the equipment racks; and a “hot aisle” in back where hot air is exhausted. To reduce cooling energy consumption, it is essential to prevent the hot and cold air from mixing. After identification of the hot and cold air flows, cost-effective measures like sealing sheet metal and hanging plastic curtains can be implemented. Other simple measures could be installing blanking panels in empty rack slots and tightly sealing gaps in and around machine rows. 
Use variable frequency drives and electronically commutated fans
Most organizations make less use of their servers at night and on weekends than they do during business hours. Yet their air handling system distributes cool air at precisely the same rate all week long. Variable frequency drives save energy by enabling air handling systems to run slower when servers require less cooling and faster when workloads are at their peak. Similarly, electronically commutated fans use “intelligent” motors to run faster or slower as needed based on airflow demand. Both technologies often reduce energy bills enough to pay for themselves within a few years.

Measures to improve MRR- Recycle E-waste and use material with high recycled content

Out of the metrics for measuring the Environmental Impact of Data Centers, perhaps the easiest to control is the Material Recycling Ratio (MRR). Simply by ensuring that all the IT equipment is handed over to authorized e-waste recyclers, and by using equipment as well as other material which has high recycled content, organizations can achieve a high MRR. 

Measures to improve GEC- Use Renewable Energy to Power Data Centers

The only way to increase the Green Energy Coefficient of a data center is to use renewable energy. Many data centers globally are choosing to install solar PV panels on their roofs, or wind turbines on the premises if renewable energy generation potential is high. 
One way to overcome the problem of frequent power outages in India could be to install solar panels on the roof of the data center. This can have backup storage and that would mean UPS of optimum capacity may be used. 

The Road Ahead is paved with Innovations

Software and hardware companies have been leading the charge when it comes to innovations to reduce energy consumption of Data Centers. These companies often tend to have some of the best technological know-how and are pursuing innovative research to solve the issues of high energy consumption in Data Centers. One of the key areas of focus has been the concept of using water-cooling, instead of air-cooling for computer chips. This would allow the ability to pack far more computing density for a given area, which can lead to massive savings in real estate space and energy savings.  In addition, water-cooling at a chip level eliminates the need for air conditioning at a data center level, and substantial savings, especially in hotter climates like India.  

Recently, IBM has been able to make a major advancement in this field. Their unique water cooling modules are embedded with the heat sinks of semiconductor chips to allow direct heat extraction from the chips to the water.   Even though this technology is not yet being deployed anywhere yet, it may see market introduction in the very near future and could radically alter the energy consumption of data centers world over. 
  4. Benchmarking Energy Consumption in Indian Data Centers, LBNL, 2010
  5. Considering an annual average household energy consumption of 1080 kWh as found in World Bank, LBNL “Urban household energy use in India: efficiency and policy implications”, 2008
  7. Benchmarking Energy Consumption in Indian Data Centers, LBNL, 2010


Author: Sustainability Outlook