It's been almost three weeks since we've announced our intention to start releasing our products as open source. I had finished my previous post by hinting that RapidCMDB, a federated CMDB solution was going to be the first open source project to be released. So what happened in the last three weeks? What's the hold up?
The delay is primarily due to a major architectural decision we have taken. All our previous products share a common server platform called RapidServer. RapidServer is all java application server that includes all the needed components to run our applications, and not having external dependencies mean quick and easy deployment.
We had initially thought building RapidCMDB on RapidServer as well, however the more we've looked at the Grails project, the more we liked what we see and started debating whether to develop RapidCMDB over Grails, rather than our homegrown RapidServer. I believe this is a common predicament for many projects. Developing and using a homegrown platform gives us full control over the platform and ensures that the platform is optimized for our needs. Using an external framework however means, we can use our resources more wisely and focus on the application itself rather than building and maintaining the platform.
At the end, after long discussions, we've decided to take a leap of faith and develop RapidCMDB using Grails. RapidCMDB is not a typical "web application" and we are still not sure whether we can do accomplish all we need using Grails. Time will tell whether we've made the right choice for the long term by moving to Grails. It certainly "feels right" to leverage other open source projects, rather than wasting resources inventing something similar.
In the short term however, this move meant delays. We have to learn this new platform and investigate how to develop functionality we had with RapidServer using Grails. We're still in the beginning of this process but we're learning quickly.
The first (alpha) release is scheduled for Thursday. This release will include functioning code and enough components to be able to get a sense of what RapidCMDB is or rather, will be. We've also been working on the development website (ifountain.org) to support RapidCMDB development. Atlassian Jira for bug tracking, Confluence as the wiki for collaboration and documentation and Subversion as the source control solution have been implemented.
Looking forward to releasing RapidCMDB into the wild...
