One of the most interesting things about the recent Open Infrastructure Summit – the new name for the OpenStack users and friends conferences – in Denver was how many people wanted to know if I thought OpenStack had yet overcome its perception problem.
And a problem it is – or increasingly, it was. The amount of interest that OpenStack got when it came out was justified, given both its pedigree with NASA and Rackspace, and the opportunity it offered as an open and flexible replacement for the various proprietary web stacks in use.
However, the hype was not, because at that point it was not really ready for widespread deployment. In particular, it could be complex to integrate and operate, and it regularly received major updates. That was less of a problem for the well-resourced telcos who eagerly adopted and customised it, but more of an issue for the average enterprise who needed support from a third-party specialist to get OpenStack up and running smoothly.
Fast-forward a few years and all that has changed. As the user-base has grown (thanks to the work of the OpenStack Foundation board, pictured above), so has the pool of community knowledge and support. The software too, like open source in general, has become both more sophisticated, more mature and focused, and easier to integrate and update.
Too blasé or too sceptical?
Step outside that user community though, and two contrasting narratives can be found. Those ‘in the know’ risk being too blasé – for some of those I spoke with at Open Infrastructure Summit, it’s just plumbing and it does not of itself deliver business value. For them, it is time to forget it and focus on what can be built on top.
The other narrative, which draws upon the disappointments that followed that excess hype of the past, is that it’s still a bit weird and of questionable reliability. Some of it is the antique notion that – with the exception perhaps of server Linux – open source is for hobbyists, or the idea that, it not being the product of a company, there will be no one willing to take responsibility for it.
Just look at what most software developers put on their PowerPoint slides when they launch a new cloud-native app: it will be their compatibility with AWS, Azure and Google Cloud. Never mind that many of the world’s largest business-focused clouds are built on OpenStack. Whether it’s by accident or design, this second narrative suits the proprietary suppliers of software infrastructure very well.
The truth is that both those narratives are flawed. Sure, it’s just plumbing, but if your plumbing isn’t solid and reliable you are going to have wet floors and mould. As to whether it’s robust and enterprise-grade, the likes of T-Mobile, Walmart and Volkswagen don’t normally build businesses on flaky infrastructure – and of course there’s now no shortage of companies willing to help you build and run your own OpenStack.