Hi Matt,

Seems to me the Qlusters team is to blame for their own demise. $34M in funding over three rounds sounds like decent support from VCs. Your April article describing how they distanced themselves from openQRM illustrates how much they lost their way…especially given the fact that the product page on Qlusters.com STILL mentions openQRM. So was openQRM important to them or not?? If not, then why is it the only thing described on their product page? If it was still at the heart of Qlusters strategy, then why did they put it on SourceForge stating “Following release 3.5 – the last release from Qlusters – we hope the community will continue to evolve and develop openQRM…”.

It apears the investors likely realized the bad state of affairs, acted like business people, and made the tough choice. Making a touch choice does not mean the VCs are to blame…just that they had the guts to make a hard business decision.

[...] LinuxQuestions.org wrote an interesting post today onHere’s a quick excerpt [...]

IMO, the problem is that configuration, provisioning and orchestration (CP&O) have traditionally been a hard sell. Also, the market for those kind of products have typically been upstream and boring (I.e., the enterprise play). OpenQRM was way to early for this market. The enterprise is just warming up to the concept of open source monitoring and CP&O is 10 times harder to sell than monitoring. IBM has struggled tremendously trying to sell CP&O in the enterprise. IBM acquired a company called Think Dynamics a few years ago and they have been trying to get their enterprise customer to move from their old configuration management software (their classic Tivoli software distribution product) to the new TPM and TPO products. In the end I think when you add the fact that CP&O is a hard sell and that an open source version is at least twice has hard to sell it’s more likely that’s what led to their demise.

Ironically,C P&O is making a huge comeback. It started with Opsware and now is finding itself a strong footing in the dynamic infrastructure discussion (virtualiztion and clouds). Even IBM has figured this out by re-branding their CP&O products as the “Blue Cloud”. C P&O for virtualiztion probably didn’t give Qlusters quite enough air to keep them alive. However, I think if their investors would have hung on for a little longer they might have figured out what projects like Puppet and Capistrano are figuring out. Configuration, provisioning, and orchestration play really nice in these infrastructure configurations that some might call a cloud.

Johmwillis.com

Maybe it’s a mixture of bad business management, lacking a good business concept(based on a true open source development practice), and bad, probably marketing-driven, development?

OpenQRM wasn’t really a product with extraordinary quality - I spent days “trying” to test it, but it was really hard to get it installed, even.
By usual open source software standards, it was hard to build, hard to install, and hard to administer.
And they didn’t even sell support in Europe. As far as I know they only took very high license fees, and would only start with data centers of a large number of nodes.
All I saw did not very much look like I wanted to build an important data center on that technology.

Last but not least, the develomment wasn’t even really open. They had a version control only accessible(even for viewing!) to company internals, and would let “some part” of the sources out to the world, on irregular frequencies, instead of conducting real open source development.

From the user point of view, a very interesting technology and product, when being shown it looked like it really would give the admin an amazing amount of functionality with some simple clicks.

Let’s see what happens with the new generation development, which is hopefully really open source happening on sourceforge…

Thanks all for your comments. Some interesting perspectives. Clearly there were a number of issues working against the company/project. I’ve written before that I think it’s as important to learns from open source’s failures as its successes.

The comments posted by ‘henning’ aren’t too far off base save one. We did manage the project correctly in the beginning, and it was the change in project management (getting away from true open source) that led me to make a career change. The efforts of the original crew (including Fred Gallagher and Matt Rechenburg specifically) shouldn’t be discounted just because the management team focused on too lofty of goals, and got away from open source (instead of looking to the community for help) when times got tough. In my opinion the story of openQRM is just now starting thanks to the dedication of guys like Matt. I’m happy to help them however I can and wish them continued success with the project.