Sankrant Sanu is a software entrepreneur, consultant and free-lance writer based in Seattle. He is a popular columnist on the Indo-American website Sulekha.com. His articles have also been published on other sites and newsmagazines including Next Future published by the Aurobindo Ashrama in Pondicherry and Manushi, published from Delhi. Sankrant has extensive management experience in the corporate world. He worked in Microsoft Corporation for over 9 years in various senior management and development roles before starting and leading his own software company in 2000. Apart from writing, Sankrant has been involved in various causes relating to India. Sankrant spent a summer traveling in rural India doing a survey of rural children to investigate issues and opportunities for rural education. His article critiquing Microsoft Encarta's entry on Hinduism led to Microsoft replacing the article with a more accurate and sensitive portrayal. Sankrant is a graduate of IIT Kanpur and the University of Texas at Austin. Sankrant has been a charter member of TIE (The Indus Entrepreneurs) and is a member of the advisory board of Infinity Foundation.

Who we are: some new reflections on the Indian experience

Dr. Krishna Palem

The unprecedented growth in the information technology industry has been significantly fueled by sustained and dramatic improvements along various technology dimension, combined with the ability to market products at relatively low costs by leveraging volume---in the hardware domain, attributes that are often summarized as versions of Moore's law named for Intel's co-founder Gordon Moore. For this growth to be commercially attractive so that industries can blossom and sustain investments, two hidden elements play a dramatic role and pose serious hurdles: non-recurring engineering cost (NRE) and time-to-market. In the ``general purpose'' computing market, these hurdles have been historically overcome on the one hand through amortizing the NRE over large volumes of microprocssors for example such that the per-chip cost os acceptable to the end customer, and on the other hand through signifcant degrees of automation embodied in EDA tools that enable the management of design complexity while accelerating design cycles and hence time-to-market. In order to achieve this, the platform that is designed and delivered is by necessity a general purpose platform that is equally good in a broad range of applicatoin contexts. In stark contrast, embedded computing applications with graphics- and speech-intensive as well as wireless contexts being representative examples---increasingly significant drivers in sustaining the growth of the computing industry---often have characteristics and demads that are not well-suited to be deployed on general purpose computing platforms. Invariably, they need some degree of customization, where NRE and time-to-market become serious impediments, except in high-volume contexts, or where the (prohibitively high) cost to develop the solution is acceptable even when only a small number of plaforms are designed and deployed. In order to ameliorate this situation and deliver the full promise of Moore's law to embedded domains with stringent customization demands without a dependence on high volumes, an entirely new generation of technologies are emerging: from instruction set extensions to tools and technologies for further compressing NRE and time-to-market. In this talk, we will survey a range of these emerging technologies and the concomitant commercial landscape. In this context, a common theme worth discussing is the increasing and necessary melding of compiler-centric design processes, traditionally viewed as being relevant in the software domain, with EDA-centric design methodologies ubiqutous to hardware design. Thus, as this trend persists, the design of embedded platforms will increasingly rely on seamless design methodolgies that will combine compiler-centric optimizations and innovatoins with EDA tools to enable solutions in contexts where the per-platform volumes are modest, but the needs for customization are very high.

Ramesh Bhashyam CTO, Teradata India R & D Centre
Ramesh Bhashyam is the CTO for Teradata India R&D center. Prior to that for five years he was the CTO for Teradata Data Base Engineering Group in El Segundo California. Teradata DBS Engineering is the primary group responsible for developing the Teradata Data Base Software. Teradata Database software is the core product from Teradata Corporation on which are based all of Teradata product offerings. Ramesh joined Teradata in 1990 as a systems architect with a Masters in Computer Science from USC. Prior to joining Teradata he worked in various engineering roles in Inference Corporation, an AI product development company, based in Los Angeles and at Teledyne Controls developing distributed control systems. Ramesh has extensive experience in operating systems, distributed processing systems, performance analysis and data base management systems. Ramesh has numerous patents to his credit and has presented frequently in leading technology forums

A data warehouse integrates large volumes of detailed and current data across entire organizations and enables different forms of decision making from the same data base. It provides a unified view of operational and historical data. It will often use as a foundation a detailed data model and enable different summaries or views, as appropriate to the business, to be built on top of the detailed model. The usage of a data warehouse has evolved from reporting and decision support system to mission critical decision-making operational systems. Data warehouses are often used for mining types of applications. These applications read massive volumes of data from within the data warehouse and are demanding of both CPU and IO resources. Data warehouses can also be used for operational decision-making applications, applications which read a small but well-focused set of data from the warehouse, and use few resources. These operational applications are differentiated by short response time requirements, making scalability a challenge when implemented on the same platform as the mining application. Data warehouses that combine both these types of applications require that the operational data be integrated, current and up to date. Further the common belief that warehouse data is static is no longer valid. There is no longer a debate that the only viable solution for large warehouses is a message passing shared nothing MPP architecture. But there are a number of challenges in implementing such architectures. The system should provide a single view for hardware monitoring, maintenance, and operations. It should provide a single system image for DBMS users and for software maintenance. From a database developers perspective an MPP platform provides a number of implementation and infrastructure challenges ^S How do you linearly scale the performance as you add mode compute power (double the compute power double the performance), how do you improve availability when the large number of components means lower MTBF, what are the challenges the optimizer must resolve in an MPP system, and how do you scale short queries. This talk will address some of these questions using the Teradata DBMS technology as the foundation. Teradata DBMS is the only large scale MPP system successfully proven in the market. Lessons from this system will be used as the basis for this talk.

Dr. Madhu Mehta Chief Architect, NirmaLabs
Dr. Mehta is a technocrat, and an entrepreneur who has been involved in innovative applications of technology to make useful products. He was an early player in software exports and exported telecom software to USA in late seventies. During those years he had represented computer industry in shaping Government policies at the national level. He consulted with C-Dot to architect telecom switching system software and to evolve software practices for project management and release control. At Anjaleem, he created attended PCO, an example of a real time embedded application, and shaped the birth of STD PCO markets in India. Later Anjaleem won national awards for export of PC based call metering systems to Saudi Arabia. Currently, besides exporting telecom call metering systems to Egypt and other countries, Dr. Mehta is setting up NirmaLabs which is a high tech incubator to spawn global techno entrepreneurs. Dr. Mehta is very active in IT Forum of Baroda and CII Gujarat.

For India to make headway in knowledge based wealth generation, it is important that students of premier institutes consider becoming high tech entrepreneurs as an alternative to corporate careers. Most of the students are unaware of the challenges and rewards of such a career choice. NirmaLabs is based on the premise that if India's technical people could be leaders in setting up high tech ventures abroad, we should be able to create a similar ecosystem to provide the missing elements to nurture such enterprising individuals to spawn what we call global techno entrepreneurs. NirmaLabs is a high tech incubator set up by Nirma Education and Research Foundation (NERF) on the Nirma University Campus at Ahmedabad. We started our operations last year in June with grooming of the first batch of participants and expect two/three â^À^Øincubatableâ^À^Ù projects by next month. The talk will initiate a discussion on the barriers to high tech entrepreneurship in India, move on to reviewing the â^À^Üseven chasms that high tech entrepreneurs have to crossâ^À^Ý to reach the â^À^Ürichesâ^À^Ý, and then outline how NirmaLabs addresses these needs to enable participants to become global techno entrepreneurs.

Prof. Naresh Dadhich Director Inter-university centre for Astronomy and Astrophysics(IUCAA)
Born on September 1, 1944 in the small village of Sarsali in Churu District of Rajasthan, Naresh Dadhich studied at various places, Pilani, Hissar, Vallabh-Vidyanagar and finally at Pune where he did his Ph.D. in General Relativity under the guidance of the doyen of Indian relativists, Professor V. V. Narlikar. Subsequently, he joined the faculty of Mathematics at the Pune University where he continued until 10th Feb. 1988 when he was invited to initiate the setting up of IUCAA as its Projector Coordinator. He became the director of IUCAA in July 2003. He is on the visiting faculty of the University of Natal at Durhan, South Africa and has also ongoing collaboration with gravity research groups in Portsmouth, UK and Billbao, Spain. Prof. Dadhich is an active and accomplished theoretical physicist of repute and works in the area of classical and quantum gravity and relativistic astrophysics. Along with his several collaborators he has published over 100 papers in internationally reputed journals. He has also trained several Ph.D. students. Apart from his academic pursuits, Prof. Dadhich has serious social and environmental concerns and has been associated with various social action groups.

I will attempt to demonstrate in a straightforward and logical manner that it was the consistency of concept and principle which naturally leads to the Einsteininan relativistic world. Further I wish to conjecture that if this line of argument was followed prior to Maxwell's electrodynamics, it would have been perhaps possible to predict it all.