The Hindu Business Line : IC to PC to GC - Evolving path of computing power

Many bioinformatics projects need high computing power that is hardware-intensive. This throws up the question of affordability. Grid computing could be a solution.

Thiruvananthapuram , March 17

FIGURE out a vast, amorphous, virtual computing resource that users could plug into and use as per needs. Well, this is as best as one can get close to distributed computing networks that deliver computing power "on tap".

IC (integrate circuit) to PC (personal computer) to Internet to GC (grid computing) has been quite a logical progression. The aim is to make computing power as easily accessible as electricity by replicating centralised generation and distribution through the grid.

The model is that of a network of computers connected together via the Internet, satellite or optical fibre links and which can be used by various organisations for sharing processing power and storage. Thousands of computers with idle CPU power morph into a single virtual computer - as in cluster grids, campus grids, renting of CPUs and storage over the Net and utility computing.

However, appropriate standards are yet to be evolved to facilitate seamless and trouble-free functioning of this grid, said Mr N.T. Nair, former chairman of the Institution of Engineers, Kerala State Centre, in a presentation made here.

Areas of application: R&D efforts and engineering designs broadly represent typical areas where grid computing is employed. Data and computing power are shared this way to accelerate and enhance design processes for scientific research and product design.

For business processes, it enables faster and effective planning and analysis of business activities. In enterprises, it helps optimise computing and data assets to improve utilisation, efficiency and business processes. It helps create low cost IT-infrastructure for e-governance initiatives.

Some typical application areas include bioinformatics, climate modelling, computational fluid dynamics, space science, high end storage services such as digital services, e-learning and GIS-based spatial information services.

Many bioinformatics projects need high computing power that is hardware-intensive. This throws up the question of affordability. Grid computing could be a solution.

Grid is also ideal for countries such as India with low PC penetration and widespread use of pirated software. A user has a low-cost box that can access, for a small fee, a central server for using an application and also for storing his files. He pays for the use only.

Broadband or LAN (local area network) or WAN (wide area network) can serve as the pipe. The need for coughing up upgradation expenses every alternate year can also be obviated in this manner.

HPC power: High performance computing through grid computing route is emerging as an alternative to supercomputers and is reaching out to mainstream businesses also.

For instance, Tata Motors, Bhabha Atomic Research Centre, Defence Research and Development Organisation, IIT-Madras and Indian Institute of Science have turned to grid computing in a big way.

Over 1.5 million PCs in 200 plus countries are now estimated to be working in grids. China and Korea have large HPC projects with 1024 Itanium 2 and 1024 Xeon processors.

In India too, there is a growing need for high-end computational resources. The IGrid initiative of the Centre for Development for Advanced Computing is one of India's grid computing projects of which Intel is a participant.

IGrid will consist of supercomputing infrastructure across cities, linked together to give an aggregate of 10 teraflops.

The grid will have 100 terabytes of storage, with Bharat Sanchar Nigam Ltd providing the optical fibre broadband connectivity. Other major players include Tata Consultancy Services, Infosys, Satyam, Cadence, Synopsis and Altair Engineering.

Currently, problems being tackled by grid computing have one thing in common: They consist of a big single problem that could be broken down into many parts and solved by sheer number-crunching effort.

But most businesses do not just have a single big problem, but many small problems in human resources to finance to supply chain management. These are to be delivered in time, unlike academic projects.

The transition from the academic realm to on-demand computing is a proving to be a tough proposition.

Article E-Mail :: Comment :: Syndication :: Printer Friendly Page

Stories in this Section
HP targets SMB storage market

Sania Mirza becomes Tata Indicom brand ambassador

TRAI explores ADC from SMS, Internet — Long distance tariffs may fall further

Anil Ambani meets TRAI chief

IT's good bet for realtors across cities

UTV Software makes debut at premium

Xilinx to expand India operations

Rambus opens design centre

BT Infonet plans nodes in Chennai, Hyderabad

i-flex privileged partner

IQ Info bags Navy order

IC to PC to GC - Evolving path of computing power