Hmm, that old teacup sized flash again....
OpenStack is and always has been a blast of hot air. Yes, it sort of does what it's supposed to but with many, many issues. The true cost of supporting it is so high as to be laughable. So many people have been burned by their support for the Project that it's hard to understand why others continue to.
Many people want to own their own infrastructure and will always want to play with the code and hug their trees. The reality however is that many organizations including said financial have been burned and realized that public cloud is the way forward. Unfortunately one running on OpenStack simply isn't going to compare to the big three and few people have the hyperscale need for a cost effective cloud anyway.
When things started OpenStack was an infrastructure orchestration play with a working storage option and a messed up and incomplete compute layer. Heck, Nova was such utter crap even NASA abandoned it. Before the project fixed that it expanded into becoming the new virtual data center operating system. But that was before anyone even got around to working on upgrades. I laughed out loud when the engineers in Paris were bitching about the pain of upgrading a cluster with six compute nodes while a competing platform could automate tens of thousands. OpenStacks issue was that it always seemed to get to the 80% completion stage before it decided that was boring and it should add more 'cool' stuff.
Then there was the whole thing about control, and who had it and whose ass you had to kiss to get anything approved. Then the explosion of projects and the confusion over what version of each project worked with each other version of other projects. It was embarrassing for all except the engineers while liked to play.
After six years OpenStack has found its place, however, while sort of works if you don't mind sitting on it and spending ridiculous effort keeping it up, who actually cares. Apart from a few zealots, the industry has moved on and frankly most customers don't really care anymore. Oh, yes a few IT shops might as they desperately cling to the idea that owned infrastructure is still king. Other, better ways of running applications exist that frankly cost less and work better. Yes, applications, that's what actually matters. Oh yes, and businesses are realizing that the value is in the app, and not the stuff it runs on, with containers and serverless compute that stuff simply doesn't matter anymore. Get the best SLA/Price combination and run it there. So while it may finally be starting to be slightly more useful than a chocolate teapot it's relevance to current need is ebbing away.
Biting the hand that feeds IT
