Good bye Linbox
As some of you may know, I’ve been working with Linbox for the past 3 years, on and off, part time and full time, depending on the time of the year.
I won’t be working with Linbox anymore after Mandriva, which has bought Linbox back in June, decided Linbox wouldn’t go ahead in the Enterprise VoIP business.
I have been, for the past 2 years, the architect of Linbox’ VoIP solution.
This unfortunate and ill-advised decision of stopping the development of a successful product at its beginning will effectively leave the current customers out in the cold, as technical support will probably be erratic at best.
Looking back in the mirror, I’ve done a lot of things, small and big, during these 3 years:
- [2004] Developed a web-managed multi-service turn-key solution for SMBs, of which the web interface served as a prototype for the Linbox Management Console (a web-based server management framework);
- [2004] Developed the product deployment system based on FAI; still in use today, obviously updated since then;
- [2004] Wrote an AD to LDAP account (including password) sync script (customer with a special need); that resulted in me hacking OpenLDAP to port the lanman password hash support to gcrypt (Debian #245341);
- [2004] Started the hosting business;
- [2005] Started the development of the Linbox IP Telephony Solution (LIPS), based on AMPortal (now called FreePBX); ended up forking AMPortal to fix the gazillion bugs and bring it to a more usable state;
- [2005] Hacked up OpenXchange to add a click’n'dial feature to the addressbook; eventually we decided to dump OpenXchange, mainly for being an underdocumented pile of crap full of security holes (nice SQL injections, eh) and, over all, a false free software (no upgrade path, no doc, some components not released, GPL patches added to the non-free version without agreement, among other things);
- [2006] Prototyped the cluster version of the VoIP solution; this cluster version meant the solution could scale up to support thousands of users on multiple sites with redundancy and failover;
- [2006] Wrote a migration tool to migrate mails away from a FirstClass groupware to anything IMAPv4-compliant, straight out of the on-disk FirstClass storage database;
- [2007] Integrated the VoIP solution into the Linbox Management Console; after 2 years of consolidating the AMPortal web interface, I’ve finally got rid of that horrible thing;
- [2007] Integrated the cluster support infrastructure into the solution;
- [2007] Added proper faxing (over PSTN) support, at last.
And, more importantly:
- [2007] Deployed client sites without any interruption in telephone service.
I’m especially proud of that; the solution was ready for that moment, and my deployment plan proved to be the right one and flexible enough to account for everything that wasn’t ready in time (missing network infrastructure, things that had been forgotten on the client side, things that didn’t go as planned and last minute changes to accommodate users’ needs). The customers were very happy and enthusiastic with regard to their new VoIP infrastructure - so much that there was more to come.
And more to come means more business for a company that definitely needs it, and is now turning that business down.
So it’s all over now, and overall I’ve had a good time working there with the technical team. They’re all very talented, and we were 3 DDs working there - that figure is now down to 1 (yes, one, as in 3 - 2 = 1).
I’ve been able to produce a really great free (as in speech) Enterprise VoIP solution that I liked quite a lot. I had great plans for the LIPS, but couldn’t carry out much of that until it was integrated into the Linbox Management Console. In the last weeks, I’ve had confirmation that my plans were right and I was going the right way.
That’s history now, and I’m looking at what I’ll be doing next.
It’s good because I can look for a job that either involves VoIP or not; it’s good to change after a while. It sucks because looking for a job sucks.