Why OpenStack needs to find a new way forward in 2018
About this time last year, Cisco announced it was shutting down its OpenStack-based public cloud. They weren’t alone. Several smaller datacentres and cloud providers, whose architecture relied on OpenStack, either entered administration or shut up shop entirely.
These types of announcements always trigger age-old rumblings that OpenStack is on its way out. When you consider that OpenStack still powers over 60 different data centres for public cloud around the world, that seems a little premature.
But regardless of whether you’re a champion of the platform or not, we have to face the fact that there are some attractive alternatives now available on the market. Enter Azure, Google Cloud and AWS – three giant public cloud providers, which have all gained a significant slice of the market since OpenStack launched, with AWS landing itself a spot in Gartner’s ‘leader’ quadrant. Not necessarily the holy grail, but the sway of Gartner is difficult to ignore.
These alternatives to OpenStack are gaining momentum, and have a number of pros and cons that should be considered, particularly when it comes to issues such as vendor lock-in, scalability and technical support.
The vendor lock-in debate
A decision to move to one of those providers may stick in the throat of anyone who fears having their hands tied by vendor lock-in. This vendor vs open source debate has raged on for some time, with OpenStack advocates claiming you simply don’t have the same freedom when you work with providers such as AWS.
Personally, I think people get too hung up on the vendor lock-in issue. Especially when, in reality, you’ll always be tied to something — whether that’s a vendor or not. For example, your platform will often be tied to a piece of code regardless of where it’s hosted. And whenever you want to move from one platform to another, they’ll always be some work required to make that change.
Regardless of your preference however, you should always be tracking the way the market is moving — which at the moment is away from OpenStack — and be prepared to respond accordingly.
Problems for OpenStack
One of OpenStack’s main problems is that it simply can’t compete with the reach, scale and wealth of experience available when working with a provider such as AWS.
While you do get flexibility with OpenStack, you’re also dealing with lots of moving parts which require skilled engineers who understand its inner workings. Unfortunately, these people are few and far between.
Unless organisations have a highly-skilled engineering team in place, it can quickly become a minefield. It requires a lot of effort to keep stable and debug OpenStack, and the learning curve is steep.
With that in mind, investing significant amounts of capital into building out and maintaining your own kit quickly becomes unappealing, especially when compared to the opex public cloud model.
The next generation of cloud
When you look at the future of cloud, it seems even bleaker for OpenStack, especially when you look at how Kubernetes has quickly become the established orchestration tool for managing containers. That layer of abstraction means we can control our containers regardless of the environment they are hosted in – be that in Google, AWS, Azure, on-premise or a combination of all four.
In addition to this, serverless services, such as AWS’ Lambda, are also making it easier to deploy and scale applications without having to worry about the supporting infrastructure. Both these solutions are changing the cloud landscape and reducing the reasons you would maintain your own OpenStack environment.
The momentum certainly seems to be moving in the direction of public cloud. That said, I’m not joining the doomsayers – those individuals who seem to relish in the demise of any technology. I don’t believe this is the beginning of the end for OpenStack – I’ve no doubt it will find its own niche. The Global Passport, for example, is a new OpenStack initiative designed to bring disparate OpenStack clouds together and this may carve out a new path. But, nearly eight years since OpenStack was hailed as a revolution, we have to acknowledge the market has changed — and appealing alternatives now exist.