By Papri Sri Raman,
Chicago/New Delhi : In a move that will empower Indian researchers and help them connect with rest of the world, the University of Chicago (UC) and two other American universities will help Indian universities ramp up their computing ability to fulfill the scientists’ dream of sharing data with the international fraternity.
The connectivity will be via high-speed ‘cloud’ computing clusters and superfast internet.
Rob Gardner, a UC researcher, has already been working with scientists at the Mumbai-based Tata Institute of Fundamental Research (TIFR) and Delhi University to connect central processing units (CPUs) at their campuses so that underused processors at remote campuses can be used by researchers in several countries.
“The main idea is to provide a cloud-based service which connect together computing and research resources at various institutions – making it easy to share data and CPU cycles for the user while not requiring specialised knowledge from local system administrators,” Gardner told IANS.
In India, the UC has proposed to build “nation-wide, internationally connected cyber ecosystems targeted for loosely coupled and scalable scientific computing applications – taking advantage of intra- and cross-institutional resource sharing possibilities, and domain and project-based collaborative environments,” he explained.
The idea comes form professor Young-Kee Kim’s team from department of physics and Enrico Fermi Institute at the UC and involves the ATLAS experiment at the Large Hadron Collider (LHC) project at CERN (the European organisation for nuclear research) in Geneva, Switzerland.
The LHC is the world’s largest laboratory and is dedicated to the pursuit of fundamental science.
Gardner is the principal investigator of the computing centre involving the UC, Indiana University and University of Illinois for the ATLAS (a particle physics experiment to discover basic forces that shaped our Universe) programme.
The UC has opened a centre in Delhi and the faculty steering committee has proposed starting a collaboration in computing – beginning with common experiences in the US and India in ‘building and operating the distributed computing infrastructure for the LHC project at CERN’.
The collaboration would begin by studying the required trans-atlantic network technical options as needed by the next run of the LHC in early 2015. The funding may come from the National Security Agency (NSA).
According to professor Amitabha Mukherjee from Delhi University’s department of physics and astrophysics, “Our department is the biggest single group using high productivity computing systems (HPCS). Collaborating with others using such HPCSs enhances our user base of computing resources as CPU intensive jobs need such remote collaborations”.
Initially, the TIFR’s CMS group ‘Tier 2’ cluster led by Kajari Mazumdar and DU’s experimental high-energy physics research group (EHEP) – generally called the Indian Compact Muon Solenoid (CMS) group would initiate the collaboration.
The CMS project involving Indian physicists is a general purpose particle detector experiment going on at CERN.
All over the world, universities and research centres are graded by computing capability and measured as of Tier 0, Tier 1, Tier 2 and Tier 3 capacities.
TIFR is a ‘Tier 2’ cluster for CMS research while DU is a ‘Tier 3’ cluster.
“People researching in the fields of plasma physics and astro-physics as well as theoretical scientists require computing capability but do not know where to find and how to use such high-end computer clusters. The idea that computing resources can be merged has been proposed by Chicago University,” DU professor Kirti Ranjan told IANS.
According to Ranjan, this CPU-sharing resource pool need not be restricted to physics research.
“It could be interdisciplinary. People delivering cancer treatment in India could search for data on a similar case elsewhere and the treatment rendered, or history departments could share data,” he noted.
For this to actually happen, a workshop involving all the stakeholders needs to be organised, scientists feel.
Then the feasibility of expansion of the physical structure is needed.
Professor Sanjay Jain, the convener of the group managing the DU physics department’s general-use cluster computing facility – which is separate from and in addition to the ‘Tier 3’ cluster – would also be part of the collaboration effort.
“This facility is used by researchers working in high-energy physics (HEP) as well as other areas of physics. With so many processors available, the idea is to better use these common resources,” Ranjan emphasised.
Gardner feels that to make this project a success in India, “we have to look carefully at existing e-science infrastructure efforts and objectives of individual institutions and departments and national efforts”.
(Papri Sri Raman can be contacted at [email protected] and [email protected])