Towards a greener planet

DQW Bureau
20 Feb 2010


With the Shared University Research (SUR) Award program initiated by IBM,

India is waking up to a new dawn of meaningful industry-academia collaboration

for joint research in areas of cutting edge technologies for tomorrow. As part

of this novel initiative, IBM has recently awarded SUR grants to two of the

leading technology institutes of India for shared studies in the areas of

developing technologies for smarter planet.

According to KPMG, the total global ICT industry is responsible for

approximately 2 percent of the worldwide carbon emissions, equivalent to

emissions from the airline industry. By 2020, even with massive improvements in

technology for energy efficiency, the IT industry will emit 3 percent of the

world's carbon emissions. It is also predicted that there will be a 50-100

percent increase in electrical costs between 2008-12 multiplying energy cost

that will amount to 50 percent of corporate IT budgets in the coming years. The

demand from storage and servers is going to increase by 30 percent, driving

energy costs further up. Having said that, it is a matter of concern that only

1.7 percent of power that reaches data center servers is used for productive


Reduction in energy and power are the key points to ensure sustainable growth

for both the ICT sector and meeting the country's bigger objectives to prevent

global warming. "IBM has invested over $1 bn to the Big Green Innovation

initiative launched in 2007. The project focuses on intelligent energy (smart

grids and alternative energy), carbon management, water management, and

computational modelling," informs Amol Mahamuni, program director, university

relations, IBM.


Moreover, IBM has announced its SUR awards for IIT Delhi and IIT Roorkee to

promote research in areas of mutual interest. The SUR awards will drive key

projects across these institutes and strive to connect research and researchers

at the universities with IBM experts in similar fields of work. A SUR grant has

been approved by the WW University Relations team for IIT Delhi. Under this

grant, a project titled 'System Level Optimization for Energy Efficient

Computing' has been agreed and will be started shortly between IBM's Center for

Advanced Studies (CAS) at the India Software Lab (ISL) and the institute. The

research project with IIT Roorkee on 'Center of Research in Data Mining' will

help in understanding the behavior of earth's ecosystem, detecting changing

patterns in the forest cover of India, and predicting natural disasters.

Smart Energy

In the context of the Shared University Research award that has been given

to IIT Delhi, the study aims to look at system level optimization for energy

efficient computing. Memory is a very essential part of energy management and in

this project with IBM is focusing at making the systems more memory aware as

part of the optimization process. "IBM has been focusing on energy efficient

computing. Not only power generation and distribution, IBM is also looking at

smart utilization of power as well. This collaborative research activity with

IIT Delhi is part of this objective," says Mahamuni.

This study on energy efficient computing essentially targets computer systems

at a larger scale and tries to device ways to bring in power efficiency at the

micro level computer and processor architecture. "As part of our studies, we

have noticed that there is a significant scope of saving energy on a massive

scale when it comes to server computers or data centers," says Preeti Ranjan

Panda, professor, IIT Delhi under whom this research work has been initiated.

"This study focuses to device techniques for reducing energy consumption of

these systems and ensure equal and efficient distribution of power," he adds.


Mapping the Data

The research activities on data mining that has been planned by IIT Roorkee

aims to establish a data mining research lab in the department of electronics

and computer engineering at the institute wherein research on data mining,

business analytics, particle analysis, and analytics laid business making will

be done. Mapping the data is essential to understand which are the actual data

that are required to be stored. At IIT Roorkee the study will be conducted on

time series data analysis from a smarter planet perspective. "We are looking at

understanding the behavior of earth's ecosystem, detecting patterns in changes

in the forest cover in India, and prediction of natural disasters," informs

Mahamuni. In addition, the research also aims at understanding the semantic

relation between unstructured data like email or text messages, blurbs,

biometric data. "We are trying to establish some hybrid techniques to do true

and efficient data mining to discover the semantic relation between these

unstructured data," he says. The research also focuses on developing new data

mining techniques for unstructured data created out of applications like email,

text messages or any other web content.

"Time series data are the data collected over periods of time that are

generated by various applications used across the world. Meteorological data is

one such example and our study will focus on that," informs Durga Toshniwal,

professor, IIT Roorkee. "Other applications include stock market analysis,

analysing customer buying patterns, identifying usage patterns of credit cards,

predicting analysis of network traffics, finding usage patterns of utility

services-all these are time series data and are research subjects for us as

well," she adds.

Meteorological data is of great importance in terms of global warming and

finding ways towards developing a smarter planet. Through this research, IIT

Roorkee is trying to use the existing data; analyze it in developing new,

efficient and scalable data mining techniques and algorithms for analyzing time

series data.

Source: DQ