In 
One has seen that the use of open source applications in large enterprises has been increasing over the years. But there is a marked divide on the kind of organisations that venture into open source area.
Mostly, the adopters are the educational institutes and the R&D organisations. Commercial enterprises still give open source a wide berth. And the reasons are clear: enterprises still continue to face challenges in using open source applications.
- Lack of support: No commitments are possible since support is usually on a best-effort basis, and that too by a set of invisible virtual people.
- Lack of a single point of responsibility: Since its not purchased, there is no commercial contract or a transfer of risk to the vendor.
- Lack of continuity: Open source initiatives are usually driven by innovative individuals in the IT department. When they leave, they take their expertise with them leaving the organisations at a disadvantage.
However, despite these factors, its still possible to use Open Source applications successfully if enterprises and IT managers use a framework approach to select applications.
We, at Sterlite, have been using open source applications quite extensively for the past three years. At present, our file servers, Internet acceleration servers, security firewalls, mail servers, document management systems, project management systems, storage servers, and desktop-level backup applications run on open source. Here, we are not talking of a single location, but multiple, large manufacturing locations on multiple servers.
And, we are not running them because they are free. Thats a misconception that many people have. The TCO might be less but its certainly not free. The kind of uptime and performance that we get is, if not better, definitely as good as provided by the commercial applications.
When you run a search, for any single requirement, one will come up with many applications. Logically, as the next step we test them out in a development environment and check out the different features and their technical fitment with the existing architecture. And then in the end, we subject the application to the enterprise litmus test for open source.
This is the clincher that helps us take the final go or no-go decision, particularly because unlike commercial applications one doesnt have a support agreement to seek help from the vendor. So it makes sense to check that the application ecosystem is sound before you commit your enterprise to it.
Development activity: As part of this criterion, one needs to check how frequently updates have been coming and what the open bugs or feature requests are. Very frequent updates or very few updates are both problematic.
Frequent updates point to an unstable system that requires frequent tweaking whereas delayed updates point to a dead or dying development cycle. The sweet spot, from our experience, is around two to three updates every year.
Add new comment