Flashback 2001: China does not have a single supercomputer in top 500 list. Cut to 2018- China dominates the planet’s highest performance machines list with 206 out of the top 500 fastest supercomputers in the world, exceeding USA (which has 124).
World’s fastest supercomputer (till May 2018) was Taihu-Light. This supercomputer owes its origin to China and functions on locally developed components.
While India and China, both started their supercomputing journey in the late 80s/early 90s, India seems to have lagged eons behind China in terms of development and installation of supercomputers.
What caused this massive traction and velocity in supercomputing development in China?
Three major drivers seem to have contributed to the supercomputing velocity that China has witnessed over the past decade or so.
Focus on Indigenous companies to develop components of supercomputers
Investments in R&D
Focus on security & proprietary
China started with off-the-shelf processors like Intel and nVidia to develop its initial supercomputers, but it realised very soon the potential possibility of technology embargo by US and started developing their own processors and other components very quickly.
In fact Chinese started using their own components and processors as early as 2011-12.
Note that India faced embargo on supercomputing technology more than a decade earlier but it hardly displayed an impetus to drive the local manufacturing of components. Even after a decade India still seems to be relying on MNCs like HP, Dell, Acer etc. to power it’s supercomputing infrastructure (Source: Economic Times, July 23rd 2018)
India, with local and indigenous focus, can not only boost it’s chances of becoming a top contender as a major supercomputing manufacturing hub in the region (look at how ISRO with proper focus, has become a major hub for satellite projection in the region) but can also look to cut the cost of acquisition and also operation with much cheaper maintenance and upgrade contracts.
The second area that India has to focus upon will be the investments in R&D. India’s USD67B is a tiny spec when it is compared against China’s USD452B in terms of R&D investments. (Source: Wikipedia). The investments can be easily built up in India too, but it seems there are not enough data points to prove the RoIs and impact on critical needs (like healthcare, security etc.) and hence there is an urgent need for multiple installations of supercomputing across the board in India, so that a visibility of long term benefits and applications is created.
The third and very critical area is security and proprietary. India should realise that they have faced an embargo of over a decade on supercomputing by the USA and other countries and hence should be wary of any such future events and should also aim to displace the dependence on brands from USA or China from both security and cost reasons.
While India is now witnessing a revival of the focus on supercomputing it must and should drive this initiative as much as possible through the help of indigenous organisations, after all it is not just a matter of security, cost and potential business boost but also of national pride.