Good day everyone. I'm pleased to post here a message from our very own Don MacVittie, outlining (literally) his research into Open Source as a replacement for that awful, "proprietary" stuff. Enjoy.
Not too long ago, a midsize corporation asked me to lend them a hand evaluating Open Source and Open Standards alternatives to their current proprietary client/server environment.
Since there's not a lot available about this type of research, we decided to 'blog it for you, so you can reap the benefits of our labor.
The scope of the research is pretty simple -- given a three-year window, does moving from the existing environment to an Open environment offer a savings?
This is a little vague, so the first thing we (there are three of us -- myself, one of the company's network admins and one of the company's client/server developers) did was attempt to put some real requirements and scope around this statement.
The applications being targeted include internal, low-to-mid-volume Web-based applications and integration applications. Nothing exposed on the Internet and nothing with a higher average volume than several hundred users per hour.
So, our research will encompass the following:
- We are looking to replace the entire Web development architecture
- Web servers
- Application servers
- Web development tools (programmer and designer)
- The proposed solution(s) must be operating system independent
- Must support Windows and Linux at a minimum
- Must be standards compliant
- SOAP
- HTML
- Java2
- XML
- WSDL
- UDDI -- possibly. We're not seeing a need at this time.
- The proposed solution must have some ease of use. We will evaluate:
- Development tool integration with Web server
- Development tool integration with app server
- Support of commonly used languages and scripting:
+ Java
+ JavaScript
+ Perl
+ PHP
+ CSS
+ HTML
- Object deployment and reuse:
+ Web services
+ J2EE
- Completeness of coverage -- what holes are there that the company does not have today?
- Since they can be handled at either the OS or the hardware level, high-availability and clustering are not required. Also, since the applications are neither high-volume nor 24x7, this requirement is lower anyway.
- To evaluate costs over three years we will include:
- Initial software investment
- Annual maintenance
- Training:
+ Application developers
+ Support staff
+ Web page designers
Next up: Tracking down the products and figuring out the "street" costs.
--Don