According to a new article in the Business Standard, Open Source software can save India $2 BN per year. Based on my experiences and discussions with Indian IT executives, that number is both accurate and low.
For the past year I have written, talked, and blogged about my own estimates (based on industry studies) concluding that of the $3.4T USD spent on ICT globally, $1T USD/year is wasted due to “bad software”. In today’s time, that’s one trillion dollars a year more than we can afford to waste. How does India, which has such a large and growing IT industry, conclude that they are wasting such a tiny fraction of this amount?
For starters, the article looks at only cutting a fraction of the waste of the waste. That is to say, rather than looking at the all-in number of just how much money is wasted on applications that are abandoned before ever going into production, or the money wasted on applications that did go into production, albeit late, broken, or missing key functionality, the study looks only at the raw costs of money being spent on inferior proprietary software that costs more to acquire than open source. Moreover, the article imagines that such cost-shifting can be done maximally at the 50% level, which is very conservative based on what I have seen from the Red Hat perspective. Today it really is possible to run open source across the entire enterprise infrastructure, from commodity servers to mainframe computers. To its credit, the article does identify an area of spending that is largely absent from my own analysis, and that is the cost of anti-virus software which is completely superfluous given the security design of SE Linux. (Note: to be secure, any system must be properly designed, implemented, and maintained: just because SE Linux may be available does not mean a system is therefore secure. But equally: with a properly designed, implemented, and maintained SE Linux-based system, the threat of a 100M PC botnet is currently nil.)
In my estimation, which is globally consistent and equally applicable across the public and private sector, the current industry practice wastes at least 30% of its IT budget on failed or under-performing applications, and until this level of quality is improved, IT will continue to be treated as a low-performer in the business. To cite one specific example from the US public sector, at the Gov 2.0 Summit in Washington DC earlier this month, Ellen Miller, the executive director of the Sunlight Foundation interviewed Vivek Kundra, the US Federal Government CIO. She pointed out that USASpending.gov was built on top of the Sunlight Foundation’s Fedspending.org database, which cost Sunlight USD $0.32M to build, yet the federal appropriate to basically provide the same data was USD $12M (it starts at 03:35 in the video). It is difficult to see these two numbers at the same time and not imagine there is some waste in the latter, despite how pro Open Source Kundra and his organization are. (Note: the CIO of one major federal department told me that the #1 value leak in his multi-billion dollar IT budget all tied back to the fact of the Microsoft Windows desktop, which had stifled innovation and resisted all efforts at proper remediation. Perhaps the $12M appropriation is due to the excess costs of supporting such legacy systems?)
I am glad to see a mainstream business resource like the Business Standard publishing a reasoned analysis and properly recognizing just how much waste there is in India’s current procurement practices. The bad news is that the numbers are likely much, much larger. The good news is that that waste can be eliminated, and India could be much, much richer the sooner it makes a more decisive shift from proprietary dependency to open source independence.