Print

Comment

Email

Digg

Del.icio.us

Reddit

Twitter

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% 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% of the worlds carbon emissions. It is also predicted that there will be a 50-100% increase in electrical costs between 2008-12 multiplying energy cost that will amount to 50% of corporate IT budgets in the coming years. The demand from storage and servers is going to increase by 30%, driving energy costs further up. Having said that, it is a matter of concern that only 1.7% of power that reaches data center servers is used for productive work.

Reduction in energy and power are the key points to ensure sustainable growth for both the ICT sector and meeting the countrys bigger objectives to prevent global warming. "IBM is committed to develop technologies and solutions for a smarter and cleaner planet. 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 IBMs 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 earths 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. In a nutshell, the project tries to identify memory mapping strategies for reducing energy. "Looking at how the systems have evolved over the years, IBM has been focusing on energy efficient computing. IBM is invariably looking at aspects like smart energy management, not just power generation and distribution but smart utilization of power as well," says Mahamuni. "This collaborative research activity with IIT Delhi is part of this objective," he adds.

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. "These systems consume huge power, most of which remains unutilized. This study focuses to device techniques for reducing energy consumption of these systems and ensure equal and efficient distribution of power," he adds. Energy efficient computing is the need of the hour and most of the IT companies are focusing towards developing technologies on that. This study is aimed to contribute significantly towards this initiative and develop newer techniques for the same. The key focus area of this research work is to understand the power load on various servers and computers in a data center and device techniques and system configurations to bring uniformity in the distribution of power that will allow huge energy saving.

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. With digitization, huge amount of data is generated regularly across the world and is estimated at 15 petabyte. 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 earths 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. Apart from time series data 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.

"The key research area under this collaboration is studying time series data mining techniques. 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 servicesall 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. "We are also analyzing the earth science data, mainly of the Indian subcontinent, for understanding the behavior of earths ecosystem, detecting pattern of changes in forest cover in India, finding patterns of changes in rainfall or the intensity of rainfall, correlation between temperature and rainfall, green hub gas emission, among others," says Toshniwal. These findings can prove to be extremely vital towards detecting natural disasters in advance, so we can figure out measures to protect our resources.

Piyali Guha
piyalig@cybermedia.co.in

Page(s)   1