There is no free lunch

Its true that Linux has been growing in prominence as pointed by Mahendran. But lets delve a bit deeper into the actual scenario. When we say "linux" we only refer to the operating system. But at an enterprise level OS is just a minuscule investment. OS licensing costs are a relatively small portion of calculating the TCO (Total Cost of Ownership) of any system. Since software, including the operating system, only accounts for around 10 percent of the total cost, reducing the direct operating system costs will only have a marginal affect on the overall IT budget. The real money-eater is Administration, development and support and its related hidden costs which account for 50 percent to 70 percent of the enterprise system. Further, there are umpteen niches that enterprises typically could encompass – Telecom, EAI, Web Services, Mobile computing, Information sharing, Migration and the list never ends. There are numerous products which cater to each of these niches. Though operating system is a vital part, these products form the basic back bone for any enterprise computing. But, more than majority of such products are not available for linux. Forget Microsoft. IBM touts linux in a big way, but they have don’t have their WebSphere suite (their linchpin product) supporting linux. Though IBM has been releasing linux based server systems, they are yet to port most of their product suite to linux. Various WebSphere suites are available for Mac OS, Mainframes, Windows and Solaris platforms but not linux. In similar lines, TIBCO, the market leader in EAI products, supports only Mac and Windows and have not indicated migration to linux as their strategy for the near future. Various SAP suites are available for Windows, Mac and Solaris but not yet on linux. With linux into the picture, integration with legacy systems becomes next only to impossible. Typical integration middleware like MQ Series do not have a linux version. Bringing out a linux version for all these products (and there are many many more, that I have not mentioned) will involve enormous investments and laborious efforts. Further, the existing infrastructure also needs to be maintained and upgraded, which makes the task all the more difficult. Linux may be an obvious choice for small Govt projects (like India’s own bhoomi, gyandoot etc.) with Java as the programming option, as they look for cheapest possible cost with reasonable quality. But considering the other wide array of product domains, linux may only be a good "Server software" and get bundled in the server systems, but not for an inter/intra enterprise-wide web/desktop/mobile application development – and this is where most of the enterprises bet their life on… So, its not a question of being free or having quality, but it’s a question of encompassing an insurmountable number of domain specific suites, which is never going to be easy for an "open source" community – the precise reason why GNU is still suffering…!!