The growth of a new technology as it enters the industry can tend to take on a certain pattern. At first, users begin to get a feel for what it does and what problems it can solve. That can lead to the next step of deploying successive generations of the technology, a greater understanding of it by customers, and more widespread use, which then drives more deployments, expanded use, and greater developer participation.
The evolution of Kubernetes, which has become the dominant container orchestration technology since being contributed to the open source community by Google several years ago, has been so fierce that it has caught up to and outran similar tools like Docker Swarm, Mesos, and OpenStack in the market. Since 2015, Kubernetes has marched through the first two phases of its evolution and is now entering its third phase, according to Brian Gracely, director of product strategy at Red Hat, which has embraced Kubernetes in its OpenShift platform.
The first phase began with Google bringing the technology behind its internal Borg cluster and container controller to the open source community, which became Kubernetes.
“We were figuring out what this community was going to look like,” Gracely told The Next Platform. “It was really early days in 2015 and 2016, with people really trying to understand what Kubernetes was, what it did, and what problems it was going to solve. In the second phase, over the next two years, we were seeing early adopters, we were seeing enterprise customers. They were using the technology, they understood containers. They were really experimenting with what applications could fit into Kubernetes and we saw quite a bit of work happening.”
Now, the technology is better understood, more types of applications are being deployed and the numbers of software distributions are growing.
“What we’re really going to see with this next jump is going to be all about automated operations around this,” he said. “People picking and choosing where they’re going to deploy this in their datacenter, in the public cloud, and kind of a mix of both of them. There will be a lot of focus on, ‘How are we going to automate the operations around this? Also, how do we better automate how to get applications onto their platforms and sort of build in lifecycle management?’”
It was with this in mind that Red Hat in January bought Linux and containers rival CoreOS and, importantly, its Kubernetes-based Tectonic orchestration system for $250 million. Red Hat had been shipping OpenShift for a few years and there was some overlap when it brought CoreOS into the fold, including between Red Hat Enterprise Linux (RHEL) and Tectonic. What CoreOS brought to the table for Red Hat was greater automation of Kubernetes operations, Gracely said, adding that OpenShift was more focused on the developer experience. In May, Red Hat rolled out a roadmap that outlined how the company was going to integrate the CoreOS technology into its own products.
Red Hat this month took the first key step in that roadmap with the release this month of the 3.11 update to OpenShift, which includes the integration of the CoreOS administration console that was a key part of the Tectonic platform and other steps aimed at day 2 operations for container support and management. The console brings a greater focus on administrators, complementing the developer focus of the current OpenShift console and offering easier management of nodes, role-based access controls and additional cloud infrastructure objects. The update also includes the CNCF project open-source Prometheus technology for cluster monitoring and for gather Kubernetes metrics and Grafana dashboards for greater visibility into and analytics around time series data.
Such capabilities are important as Kubernetes evolves. Gracely said the addition of the CoreOS technology will offer greater support for ISVs that are looking to package their software onto the platform and building in lifecycle management capabilities. Software developers are looking at Kubernetes as a platform that will enable them to write their applications once and run them in datacenters as well private, public and hybrid clouds, he said, adding, that OpenShift 3.11 is “a really good deal in terms of simplicity for ISVs, simplicity for customers, and now having to be so worried about the public cloud being different from the private cloud, or is one public cloud different from another public cloud.”
Growing OpenShift is a key part of Red Hat’s strategy going forward, particularly as renewals for RHEL appears to be decelerating. The company in September reported a 14 percent year-over-year jump in revenue, to $823 million, with guidance for the third quarter coming in at between $848 million and $856 million, which would be a similar increase. However, much of the focus of the conference call with financial analysts was around an apparently slowdown of RHEL renewals and the rapidly-growing interest in Red Hat’s OpenShift and Ansible software automation technology.
During the call CEO Jim Whitehurst said that Red Hat had picked up 100 new OpenShift customers and that in a recent trip through Europe, more than 20 of 30 customers he spoke with wanted to talk about OpenShift. Whitehurst also noted that Hortonworks and IBM Private Cloud will bring their big data analytics products to OpenShift, and said that the platform’s growing popularity is having an impact on the company’s sales staff.
“It really is a topic that people want to understand and figure out how to do in containers,” Whitehurst said on the call. “I also think one of the things that has slowed RHEL growth a bit is customers want to talk about containers, and we need to continue to talk about RHEL as part of that portfolio. As we get into next year, we have a much larger renewal base. And since OpenShift is continuing its momentum, I would say it can easily accelerate quite a bit, but we are running into a little bit of a sales capacity issue. And so, the better OpenShift does and the more time we’re spending on that, the less time we’re selling RHEL.”
Gracely said the company ships OpenShift with RHEL and that the two pull each other along in terms of technology and deals. Overall, the core RHEL business is strong, he says, noting that “we still take on new customers every quarter and every month, but a lot of that business tends to be renewals. The OpenShift business is a much broader platform and involves a lot more decision-makers because it does touch applications and it does touch infrastructure. That’s what Whitehurst was conveying, that deals for platform-level things tend to involve more people and tend to be more involved in terms of decision-makers and the process takes a bit longer. That can potentially take an account team away from their repetitive renewal business.”
Gracely notes that OpenShift has seen almost 100 percent customer growth year over year and that its presence in multiple clouds and public clouds is helping to drive RHEL into public clouds, opening up new markets for the company.
The continued integration of the CoreOS technology will help continue that push, Gracely says, adding that the integration is about 40 percent completed with OpenShift 3.11. The remaining integration will be reached with the next update early next year. It will further help customers adopt Kubernetes, which he said has been surprising in how quickly enterprises and other organizations have embraced the technology.
“We obviously like the product and like the technology, but customers have been really impressive in terms of, once they began to really understand what Kubernetes could do – it’s built-in scalability, it’s built in automation, it can work across a bunch of different applications – we’ve been surprised at how aggressively customers have been trying to put more and more onto it,” he says. “Sometimes you bring in technology that takes a while to just work out, but we’ve got customers with 15 or 16 distinct divisions running this, so it’s applicable to banking and finance, we’re seeing it automotive, we’re seeing it in transportation, we’re seeing it a lot of different verticals. For us it’s been a real big differentiator and for the company as a whole it’s been one step beyond our core Linux in our ability to help our customers with applications and to help them with broader cloud architectures.”