Cloud Foundry is not capitulating on its dream by embracing Kubernetes—it is fulfilling that dream.
Cloud Foundry might have its issues similar to another group, however arguably, its choice to embrace Kubernetes is not one in every of them, regardless of strategies on the contrary. In truth, as Cloud Foundry Basis government director Abby Kearns put it, “Our imaginative and prescient is to be the most effective cloud native platform for builders,” irrespective of the underlying applied sciences.
SEE: Cloud suppliers 2019: A purchaser’s information (free PDF) (TechRepublic)
The dying of a disco dancer
Paradoxically, whereas criticizing Pivotal for “hyper advertising and marketing” the Cloud Foundry platform, Rishidot analyst Krishnan Subramanian does just a little hyper advertising and marketing himself. On the eve of a well-attended Cloud Foundry occasion, Subramanian wrote a eulogy “on the ‘demise’ of Cloud Foundry.” I assumed I had missed a serious announcement—that the venture had one way or the other folded. I scoured the web, in search of information of this “demise.”
To no avail.
Sure, I had associates pinging me to inform me that Pivotal, the corporate behind Cloud Foundry (now housed in its personal basis), was struggling. I checked out its financials: Narrowing losses and elevated revenues. The inventory value is not at its peak, however it has been shifting up persistently during the last six months. LinkedIn exhibits that the whole head rely has declined 14% over two years, and down three% during the last six months. Not setting the world on fireplace, maybe, however not useless, both. (And, as Start co-founder Brian Leroux has noted of Pivotal’s $5 billion market cap: “Hope I fail our a—— right into a $5 billion area of interest enterprise! lmao.”)
And that is Pivotal, not Cloud Foundry. So when did Cloud Foundry die?
Digging into Subramanian’s submit, it turns into clear that he is utilizing “demise” in a really explicit approach. Noting that Cloud Foundry’s unique container format (Backyard, most not too long ago) and orchestration engine (Diego) had each been changed by business requirements (Docker and Kubernetes), he concludes: “As soon as you are taking the container runtime and container orchestration out of the equation, no matter stays from the unique Cloud Foundry platform has very restricted relevance.” Or, in his rendering, is useless.
Once I requested for clarification, he obliged: “There isn’t a magical expertise or effectivity supplied by Cloud Foundry which isn’t already accessible in Kubernetes ecosystem. This announcement solely helps current Cloud Foundry clients to make use of Kubernetes as an alternative of Diego. In any other case, Cloud Foundry as an idea is useless.”
Is it, although?
Properly, it occurs loads round right here
With one caveat (which I am going to come again to), for a corporation like Pivotal (or Pink Hat or…identify your enterprise vendor) it actually (actually) would not matter what the underlying applied sciences are. If Pink Hat powered OpenShift with magical hamsters however clients obtained the size and ease-of-use they require, nobody (besides perhaps PETA) would thoughts. In like method, the truth that Cloud Foundry has been altering the applied sciences that energy its platform is immaterial to buyer success with that platform.
SEE: Vendor comparability: Microsoft Azure, Amazon AWS, and Google Cloud (Tech Professional Analysis)
As Kearns put it to me, “For the final couple of days, I have been having conversations with customers at our summit, customers who’re operating tens of 1000’s of functions on Cloud Foundry, and plan to extend these numbers.” When Cloud Foundry started, it supplied performance that Kubernetes could not match. That has since modified, which is why Pivotal engineer Oleksandr Slynko may write in October 2017 that Kubernetes would by no means exchange Diego after which in Could 2018 he reversed his adamant assertion to say “Certain it could possibly.” Pivotal, and Cloud Foundry, needed to evolve to embrace what the market had chosen.
That call, nonetheless, is pushed by clients. As Kearns famous, “Having a secure platform is of the utmost significance. Being proper on the know-how hype cycle just isn’t of the utmost significance.” As Kubernetes matured and solidified business assist, it grew to become the plain reply to Cloud Foundry’s container orchestration wants. Over time, Kearns went on, that may effectively change, with Cloud Foundry dedicated to swapping out applied sciences to maintain with its core mission: “Our imaginative and prescient is to be the most effective cloud native platform for builders. It’s best to by no means financial institution on one particular, tiny know-how to be the tip sport. That is not the tip sport. The top sport is to have a platform that permits clients to construct, run, and scale functions.”
Or, as Pivotal vp Richard Seroter said, “Kubernetes is nice. That is why it will be a part of Cloud Foundry, as are a dozen different nice open supply software program tasks we have swapped in over time. However the worth of automating the trail to manufacturing in a SINGLE platform is effective by itself. And other people pay for packaging of that, versus DIY.”
Which brings us to the caveat I referenced above.
For those who suppose peace is a standard aim
There’s worth in embracing an business pattern first, and serving to to form it, as Pivotal/Cloud Foundry competitor Pink Hat/OpenShift has carried out with Kubernetes. Again in 2015, Pink Hat opted to skip becoming a member of the Cloud Foundry Basis, and as an alternative doubled down on Docker and, later, Kubernetes. Writing on the time, Pink Hat’s Joe Fernandes highlighted the significance of this selection:
With Linux containers on the core of many trendy cloud utility platforms, we felt it will be a misplaced alternative to not embrace an rising container normal. Pink Hat has all the time labored to drive upstream requirements within the Linux group, to assist keep away from know-how fragmentation, and with containers we see the necessity for the same strategy. Requirements for container codecs (together with packaging, APIs, index metadata) additionally allow us to collaborate with distributors like Google, Microsoft, IBM and others who’ve equally determined to assist Docker.
Related motivations underlined the choice to go together with Kubernetes:
Given Google’s deep experience in container know-how and managing what is probably going the biggest containerized datacenter infrastructure in existence, we felt that standardizing on Kubernetes can be the appropriate selection for OpenShift customers. At present Pink Hat is contributing extensively to Kubernetes, and dealing with the group to convey these capabilities to [Red Hat] clients…. And in contrast to Cloud Foundry Diego, we aren’t investing in an OpenShift-specific container orchestration resolution, however one that’s being supported by various distributors together with Microsoft, IBM, Mesosphere, CoreOS and extra.
Cloud Foundry finally reached the identical conclusion, however there’s worth for Pink Hat in having such a deep footprint within the open supply communities round Docker and Kubernetes. Whereas Pink Hat would not management both group, it has “purchased” itself affect with the code it has contributed. This places it in a stronger place to assist these still-emerging business requirements.
So, is Cloud Foundry “useless?” Nope. Not even shut. However will or not it’s that rather more alive if it, sooner or later, helps to drive business requirements like Kubernetes, as Pink Hat has carried out, reasonably than come to the celebration belatedly? Sure. Even so, full credit score ought to be given to Pivotal for reaching the appropriate conclusion in any respect: It is exhausting to alter know-how route after huge investments have been made. They’ve carried out this greater than as soon as, and that is a credit score to its buyer focus.