In what will be an unusual turn of events, I only have a short hop across the Tasman Sea to attend an OpenStack Summit. Whereas past summits have occurred in far-flung locales, this event is taking place in Sydney and so, for a change, it is my colleagues, and not I, that has the long hike to attend.
I’m really looking forward to this summit and, in particular, am looking forward to seeing how OpenStack reacts to the ever-increasing attention hat another couple of open source projects, Cloud Foundry Kubernetes, is gaining.
Some Kubernetes and Cloud Foundry data points, for reference
To be honest, there is always a fine line when it comes to industry buy-in for these open source projects. Readers may well recall that in the early days of OpenStack, seemingly every technology vendor under the sun jumped at the chance to become. But the difference is that Kubernetes seems to have attained far more end-user buy-in, early on, than OpenStack did. Fully 40% of Cloud Foundry’s membership comes from the enterprise world, which can be seen as a proxy for the fact that, rather than a project that is simply driven by vendor’s commercial imperatives, Cloud Foundry is actually seen as important to enterprises’ future strategies. According to a study by Redmonk, at least 50% of the Fortune 100 are using Cloud Foundry.
For its part, Kubernetes seems to be going from strength to strength. While far newer than Cloud Foundry, Kubernetes has had some fortunate timing. Docker popularized the idea of containers and Kubernetes was created, at least in part, to counter some of the concerns around Docker’s intentions with its own Swarm orchestration play. Since then, every vendor under the sun seems to have jumped onto the Kubernetes bandwagon and my sources working within vendors offering Kubernetes services indicate that there is a growing groundswell of interest in using Kubernetes.
Indeed, these two projects, Cloud Foundry with its PaaS-centric origins, and Kubernetes with its container-centric ones have come together somewhat. Kubernetes is a first class citizen in the Cloud Foundry ecosystem with Google and Pivotal having decided to hand over the Kubo project to the Cloud Foundry Foundation.
Of course, this post is all about the OpenStack Summit and hence too much gazing over the wall into PaaS and container land isn’t helpful, but it is certainly worth recognizing the situation OpenStack finds itself in – arguably attempting to push a construct that is going out of favor. While a notoriously bad measure of actual adoption, the Google trends plot comparing OpenStack, Cloud Foundry and Kubernetes is… interesting. Kubernetes is obviously in its ascendency and having a negative impact upon OpenStack.
Enough of that stuff, what will the OpenStack Summit bring?
The executive director of the OpenStack Summit, Jonathan Bryce, must be really getting sick of the constant claims that the initiative he has been fostering since its inception isn’t really “a thing.” Bryce is continuously having to defend OpenStack from those who say that, variously, Amazon Web Services, Docker, Containers, Kubernetes or whoever have “won.” As if this game were in any way a zero-sum one.
The reality is, as I have articulated many times before, that many large organizations are successfully using OpenStack to run their most critical workloads. Large webscale organizations such as eBay. Huge science projects such as (famously) CERN. Giant telcos on every continent. All are using OpenStack to some degree. Of course, that doesn’t mean that OpenStack has won, and I have to say that Mark Collier, OpenStack Foundation’s COO, was gilding the lily when he stated in a recent interview that:
In the infrastructure world, it’s absolutely moving to cloud in every form. In the open infrastructure world, OpenStack is the de facto standard. When you think about people innovating in data centres and at edge computing – the foremost experts in those fields will be in the room in Sydney. Time is of the essence for every company – they want to move faster. This is the best place in the world to be to talk to talk to people who have architected extremely scalable systems, the people who implemented it, the people who made the mistakes that you don’t want to repeat – learning from people who have done this cloud transformation you might be tasked with – this is where you want to be
There are many open infrastructure project contributors what would argue there are other standards. Notwithstanding the hyperbole – big organizations are doing big things with OpenStack. But the question still remains, how will OpenStack play with everything else that is going on.
Finding the intersects with the broader open source world
Collier commented on OpenStack’s need to play nicely in a bigger world in the same interview when he opined that:
The question is what’s going to be combined with OpenStack to power this big wave of infrastructure. People are doing things with Kubernetes and OpenStack or other interesting container-type projects. We’ll be hearing from them as well
So really I’m expecting some more clarity of thinking around how OpenStack plays within the broader ecosystem – particularly alongside Kubernetes and Cloud Foundry. I’m keen to see some clarity around use cases for the different solutions and some work to be done making sure that the user experience for those organizations using multi-cloud is consistent and reliable. Better integration and more robust testing will help the community build the perception that OpenStack is a logical and safe place to run these other tools.
Better interoperability
But multi-cloud isn’t just about OpenStack interfacing well with third-party platforms, it’s also about providing better interoperability between the different cloud built on top of OpenStack – we saw some moves around OpenStack interoperability at previous Summits so I’d expect to see more of that this time. After all, one of the OpenStack propositions is that it creates the ability for end users to leverage a variety of different cloud providers, all with high-level of fidelity. Delivering on that promise is critical.
Vertical clouds?
We’ve already seen OpenStack achieve good success in the public sector and across the telco and educational sectors. Maybe it’s time now that the platform is stable, for OpenStack to focus on some other verticals. I note that Australia’s Commonwealth Bank has some deep involvement at this summit – could that be an indication that a financial services focus could be coming to OpenStack?
Silencing the critics
Above all, though, beyond interoperability, commonality with other projects and other initiatives, I just want to see some really good, chunky use cases for OpenStack. The Foundation needs to find ways to, once and for all, silence those who claim that OpenStack is a failed experiment. The project has largely moved away from its early “AWS-killer” days, now is the time to double down on what OpenStack is, rather than what it isn’t. The Sydney Summit is going to be interesting.
PS
Since the summit is being held in Australia – this talk, looking at local OpenStack adoption, should be worth a look. Also, my mates at Aptira tell me that they’re having a party on a boat in Sydney harbor – I’ll be there as well so come along and say “gidday.”
