I’ve just returned home, crossing the planet after attending the OpenStack Summit in Austin, TX. Once again I was impressed at the progress made by the OpenStack community and the improvements in number and sophistication of OpenStack-based offerings. My last OpenStack Summit was Paris in late 2014, and that was a great event also. Afterwards, in early 2015, I wrote Is OpenStack Ready for Mainstream Private Cloud Adoption? that became a very popular Gartner document and summarized our position at the time.
My colleague Donna Scott was asked to present at this year’s Austin OpenStack keynote on cloud native workloads and Bimodal IT. I was sitting in the keynote when the tweets started. “Last year Gartner said
#OpenStack is a science project – this year they are part of the opening keynote,” lamented one pundit. “Gartner no longer thinks OpenStack is crap” said Boris Renski from Mirantis in his session. Another post decreed that Gartner was “now scrambling to recover after realizing their error.”
“Hang on,” I thought, “did I ever say OpenStack is crap? Have I changed from not recommending it to recommending it?”
Well here’s a great lesson in perception vs reality. I’m happy to discuss criticism of Gartner’s positions but let’s start at least with reality and not a strawman. Let’s look at the facts.
The basis of all this seems to be an article by Simon Sharwood in The Register that he wrote after attending my session at the Gartner Infrastructure, Operations and DataCenter Summit in Sydney in May 2015: OpenStack private clouds are SCIENCE PROJECTS says Gartner .
Let’s start with the offending quote in full from the article in El Reg:
“OpenStack is great as an open source standard for infrastructure access,” Waite told the Gartner IT Infrastructure, Operations & Data Center Summit in Sydney today. “It has great APIs. But it is not a cloud management tool. It is a framework on which you build and this is why people get into trouble: it is a science project and you need to be aware what you are getting into.”
That’s it. At the time one of the biggest misconceptions I found in Gartner clients was that they thought they could “download OpenStack and replace VMware”. This is around the days of Grizzly/Havana/Icehouse by the way – if you spoke to anyone building OpenStack themselves from scratch back then and trying to scale it – it certainly WAS a science project!
Now let’s look at some other quotes from that article:
“OpenStack, he predicts, will be a success, only the degree of that success is to be decided.” … “The main thread in Waite’s talk was that if you decide OpenStack really, really is for you, Don’t Try This At Home. Hire the few experienced hands on offer, brace yourself for bumps and, above all, don’t even get to the starting line unless you absolutely need infrastructure-as-a-service (IaaS)”
So in the very article with the infamous “science project” quote I actually said “OpenStack will be a success” but I suppose that headline doesn’t drive as many clicks. I said the same in my blog 12 months ago: Is OpenStack a Success? “Sure it is” was my answer.
So what is Gartner’s position on OpenStack? Well it hasn’t changed from a year ago. Here are the major documents we have published in the last 12 to 15 months:
- Assessing OpenStack’s Viability for Top Use Cases
- How to Succeed in Your OpenStack Deployment
- Insight Into OpenStack Packaging and Consumption Options
- OpenStack Is Not a Cloud Management Platform
- Is OpenStack Ready for Mainstream Private Cloud Adoption?
I’ll quote the Recommendations section of my research article (the last one in the list above) since it is behind the Gartner paywall:
- Build a private cloud with OpenStack if you require true cloud IaaS characteristics on private infrastructure, you are an organization with access to substantial engineering resources, and you are committed to open-source principles and seeking a “standard” infrastructure provisioning API, reducing lock-in to current proprietary vendors.
- Use a managed or outsourced OpenStack-based private cloud if you meet the above requirements but lack engineering skills or budget to implement and maintain the complex environment.
- Base your private cloud on a packaged OpenStack-based distribution/CMP (with support from a large vendor) if you meet the above requirements but must implement an OpenStack environment yourself and are less concerned about lock-in and more concerned with simpler implementation and support.
The key here is that OpenStack is a great solution if you fit the use cases for it. From my blog a year ago:
When Gartner clients ask me if they should be using OpenStack in their organization I first ask three questions:
- Should you be building an IaaS private cloud?
- Do you have the skills and resources available for a project of this complexity?
- Is an open source framework like OpenStack the right tool for the job?
For most organizations the answer to at least one of these questions is NO.
Now as OpenStack and its ecosystem improve, the number of users who fall into these use cases expands. Gartner is seeing increased interest in hosted and managed private clouds, for instance. The OpenStack community is starting to address the complexity and skills issues. We also see OpenStack gaining ground in smaller public and community cloud providers in geographies such as China, Middle East, and parts of Europe where the US based cloud providers are not as popular. So the ecosystem and community evolves, and so too do our positions. That’s what research analysts and media are supposed to do. What we should not do is base an opinion on someone else’s headline, and write articles without so much as reading the full source article or trying a simple Google search for sources.