| The best Zimbabwe news site on the world wide web |  |
| NEWS |
| FORUMS |
| NEWS ANALYSIS |
| READERS' FORUM |
High performance computer for bioinformatics in Zimbabwe
By Tendai
Padenga
BIOINFORMATICS is one critical area that surely needs the attention
of developing countries if any meaningful and permanent solution is
to be nurtured and realised in as far as accessibility and reduction
of bulk drug prices and job creation is concerned.
It is surely important to acknowledge that without a comprehensive road map in building a vision-based, internationally well-anchored policy on how to tap the potential that lies in this hot area, we will not benefit much from the minds of professionals like Prof Chetsanga and many more.
Bioinformatics is a very exciting and at the same time challenging area and demands the use of High Performance Computers (thereafter referred to as supercomputing facilities). Focus on this area is necessary if Zimbabwe is to domestically perform such exercises like protein folding, drug design and genome analysis.
If such a supercomputing facility is set in place and well-managed, meeting the known application loads and with the potential to adopt to the unknown application loads that exist today for upward compatibility later, we are assured of better technological research turnover.
We should not forget that today’s great demand simply lies on the ‘time to market’ of any given product and without any second thought, putting in-place a supercomputing facility for research in bioinformatics will aid Zimbabwean researchers as well as universities, those in the Diaspora and companies with interests in such areas, to start and continue collaborated research work using Zimbabwe’s own supercomputing infrastructure.
Let me be very frank here, availing supercomputing facilities to scientists is priceless, it prompts them to live their dreams and the moment they realize they can do what they can dream then they dream even bigger and that cycle will grow the industry. There is no meaningful industry, to my knowledge to date, that brought about national development and did not have its roots in research and development and the world today is demanding more complex and highly analytical research work that if not aided by the supercomputing facilities, will make research meaningless since it will not make business sense.
If we are to draw lessons from either India or Turkey’s experiences, we will notice that these countries are winning the bulk drug manufacturing 'crisis’ through groundbreaking collaborative research work in bioinformatics by simply using their own supercomputing facilities and making maximum use of them. Let me clear the air that most Indian and Turkish supercomputing facility centres I have visited have spent very affordable amounts on the same, they have not spent millions, not even USD$30 000, on these supercomputing facilities but the work that is being done especially in India has amazingly propagated development in the drug industry and to date Zimbabwe is among one of the African countries purchasing these bulk drugs from India.
The trick behind setting up a comprehensive and yet cheaper supercomputing facility is to work the architecture with a heavy bias towards component based technologies and then later do our shopping in either India or China. It would be very short sighted of us to write off Indian, Japanese or Chinese technologies. We need a simple but comprehensive plan of technology adaptation and proper understanding of the application demands at hand and then carry our technological shopping baskets to the same countries.
One suggestion that comes to mind is never to buy products from these countries especially in areas where the need of the hour would be highly application area dependent. We should only purchase those components that we critically need. Like I said, we have graced our roads pretty well with Eastern technologies -- need I mention Toyota, Mazda, LG televisions and DVD players? The same can be done easily with supercomputing facilities, the only pivotally determining factor being comprehensive knowledge of our own domain problem, otherwise the solution is ready in component form in these Eastern countries and at very reasonable prices.
Allow me to highlight in passing that a high performance computer for a developing country precisely tailored for bioinformatics, as the main application plus various core software is highly affordable for National Development and this can be set-up without any struggle provided the government ministries and the various research institutes come to the true understanding that it’s the only way forward. Any compromise on it will surely mean compromise to the ultimate future of self reliance through scientific driven research as a vehicle to propel the industry and improve the lives of Zimbabweans through employment creation and better access to basic drugs.
Having participated in the design, purchasing, assembling and operations of such supercomputing facilities, I would want to outline that the following are basic infrastructural requirements for setting a supercomputing lab which can basically get Zimbabwe on the bioinformatics research highway:
1. 80 Ultra Sparc lll Cu 900 MHz processor cluster over a Gigabit switch (i.e. 40 dual processors)
2. Two 16 node P4 processor clusters
3. A 4 processor SGI ORIGIN-200
4. Five Sun Fire V20z Server
5. Sun StorEdge 6120 Array
Ladies and gentlemen allow me to boldly say in confidence that the above infrastructure, if setup properly, from my experience, will have an aggregate computer power of around 200 Gflops and the data storage capacity can be around 10 Terabytes. This alone will create an adequate starting point on the bioinformatics revolution in Zimbabwe. Yes, it might not make much sense in general, but try asking Prof Chetsanga what he can do with that much computing power in bioinformatics.
Assuming all things are equal, the set-up of such a facility and the exchange of bioinformatics research programs between Zimbabwe and countries like India, China etc will bring in more value at lesser dollar cost, ultimately leading to a better Zimbabwe.
Tendai Padenga
is a visiting research scholar to Osmangazi University in Eskishiher,
Turkey, and has research interests in Massively Parallel Computing and
Artificial Intelligence. He can be contacted at tepadenga@yahoo.co.uk
JOIN THE DEBATE ON THIS
ARTICLE ON THE NEWZIMBABWE.COM FORUMS
debate@newzimbabwe.com
All material copyright newzimbabwe.com
Material may be published or reproduced in any form with appropriate credit to this website