Is Cloud Foundry behind the times or ahead of the game? And, perhaps more importantly, does it even matter?
Those are the two main questions I have coming out of the recent Cloud Foundry Foundation North America Summit 2019. The event itself was … well … really just another trade show that had some good insight but was held in a venue that was 2x too large. But it was that insight more than anything that has me in such a quandary.
The main takeaway from the event was tied to the notable progress the organization’s Project Eirini has made since its adoption into the Cloud Foundry Foundation last October. Cloud Foundry Foundation Executive Director Abby Kearns told attendees during a keynote address that Eirini had passed core functional tests “and is now mature enough that early adopters have begun to deploy in production environments.”
What’s Project Eirini? Basically, it’s a bridge between Cloud Foundry’s legacy Diego container management system and the Kubernetes container orchestration platform. More realistically, it’s the platform on which Cloud Foundry users will build their Kubernetes-orchestrated container futures on top of.
The move to further embrace Kubernetes has been a long-time coming. At last year’s Cloud Foundry Foundation event, organization executives were cautious on their relationship with Kubernetes. Sure, Kubernetes at that point was viewed by many as the de facto container orchestration platform, but the Cloud Foundry Foundation was not quite ready to ditch the stability of its long-serving Diego platform.
However, that tide became too much for the Cloud Foundry Foundation to hold back, and it’s now on the road to replacing Diego with Kubernetes. And make no mistake, Diego is living on borrowed time. Sure, some of its tenants might remain in the near term, but it will be replaced by Kubernetes.
Cloud Foundry Foundation CTO Chip Childers explained that this evolution is natural and something that Cloud Foundry has been through in the past when it migrated from DEA to Diego.
This move to only now fully embrace Kubernetes would make it appear that the Cloud Foundry Foundation is indeed behind the times. But, in talking with enterprises that actually have containers running in production environments, the vast majority are on board with taking a more measured approach.
Kubernetes remains a hard platform to use in production environments, which should be amply clear by the numerous managed Kubernetes platforms being offered by vendors. And the terms “difficult,” “steep learning curve,” and “challenging in production” were mentioned in numerous conversations and interviews while at the show.
But, despite those dour tones, all of those I spoke with know that Kubernetes is the future and that they need to get on board with that program. And, they are happy that Cloud Foundry is using Project Eirini as a measured approach to the journey they are only now just starting.
“We’ve been pushing for this for some time,” said Jason McGee, vice president and CTO for IBM’s Cloud Platform. “For the Cloud Foundry community this will free up some energy in the community to work on other problems. That community can now focus more on the front-end experience where I think Cloud Foundry is really good at.”
It should be noted that IBM has been one of the lead vendors on developing Project Eirini, but McGee’s sentiment was echoed by most that I spoke with at the event.
What this move will mean foundationally for Cloud Foundry is that it will become the application-control layer sitting on top of Kubernetes, which will act as the infrastructure-control layer. Some graphs at the show has Cloud Foundry occupying a layer either next too or just below the Kubernetes-based Knative project, which is likely to be the next area of focus for the Cloud Foundry Foundation going forward.
Or will it?
Cloud Foundry continues to have a dedicated and loyal group of enterprises using the platform and vendors providing support for those enterprises using the platform. But, I would also guess that the platform is not going to see a lot of growth from those two segments moving forward.
As mentioned, those vendors all have their own managed Kubernetes platforms that include different components that ride on top that provide similar functionality to what Cloud Foundry offers. And, as mentioned before, enterprises know that a Kubernetes-based management layer is the future and that they will eventually need to fully jump on board. Those two factors could marginalize Cloud Foundry’s place in the ecosystem. I am not saying it will go away completely, but it will have to evolve at a more rapid pace in order to survive or search out a new future.
Jay Piskorik, director of platform engineering at Dick’s Sporting Goods, noted during a panel session that even though his organization only recently began using Cloud Foundry, he is already planning for its evolution. He said that he expects his organization’s use of the platform will become “a flavor of Cloud Foundry.”
“We will not be talking about it in the same way,” Piskorik said. “But I am sure Cloud Foundry will be part of that.”
Troy Topnik, senior product manager for SUSE’s Cloud Application Platform, offered up a more detailed suggestion, noting that the Cloud Foundry Foundation could at some point be merged with the fellow Linux Foundation-hosted Cloud Native Computing Foundation (CNCF). That would definitely cement the Kubernetes connection into Cloud Foundry seeing that the container orchestrator is housed within CNCF.
The Cloud Foundry Foundation has indeed made some important moves over the past six months that could set it up for future success and enhance its community support. It’s now up to that community to see what will become of the organization over the next year.