IBM paid $34B for this: Where Red Hat is taking OpenShift

Orchestration gets its own conference: Why KubeCon is now a thing
Scott Fulton says thousands of IT professionals descend upon San Diego to see frank discussions and live demos. Is there a cohesive strategy for the users of, and contributors to, Kubernetes going forward? Are we ready to make service mesh a part of our lives? Read more: https://zd.net/2s948pe

IBM already had a Linux. Linux is no more the lead product of a computing company than aspirin is the lead product of a pharmaceutical company. When IBM announced in October 2018 its purchase of Red Hat, its executives explained the move using marketing-speak, calling it an investment in the hybrid cloud. Business news services duly noted that Red Hat was the producer of Red Hat Enterprise Linux (RHEL), which is generally believed to have at least a two-thirds market share in enterprise servers.

Mentioned parenthetically, if at all, was something called OpenShift. It has been Red Hat’s platform-as-a-service (PaaS) system — a way for enterprises to pool together the servers they already own, integrate resources from the public cloud, and deploy applications built for the system directly from source code. (If you hear the term “cloud-native” bandied about like an ice cream flavor, that’s what it actually means.) In the type of move a major software company typically never makes, in 2015, Red Hat abandoned the architecture it had already created for OpenShift in favor of a containerized system based around Kubernetes, the emerging orchestration platform of that time.

The unavoidable fact of our industry now is that the roles played by hardware and software have been swapped. Infrastructure today not only comprises the servers of but also the servants to, the software that comprises the world’s applications and services. To be in the server business isn’t enough; one has to have a hand in the infrastructure software that puts those servers to work. Otherwise, it wouldn’t be “Dell Technologies” now.

So it was, on an unseasonably hot mid-November day on a chartered yacht in the Bay of San Diego, that the newly inaugurated IBM corporate “unit,” assigned to its hybrid cloud division, set forth its strategy to do for enterprises in 2020 what IBM’s AS/400 Series did for them in 1988. Speaking over the clanging of the diesel generators in the engine room, and rocked by the wake of the occasionally passing shipping barge or naval destroyer, here was the strategy for which the most storied corporation in the history of computing paid $34 billion.

Daily affirmation

“Every single company at this conference is trying to frame themselves at the open-source infrastructure company for the next version of this story,” said Andrew Clay Shafer, Red Hat’s vice president of transformation, referring to the Kubernetes-oriented KubeCon conference which this meeting on the yacht was based on. “So what is the opportunity to differentiate yourself in that market, where everyone’s going to try to tell you they’re the open-source infrastructure company?”

The common belief among many business news writers is that open source is a kind of public relations tactic, a side venture where designated, non-critical practitioners can gather together at conferences like KubeCon in San Diego, and make it appear to the world as though they’re working together in the spirit of collaboration, agility, and insert other business-book metaphor here. As it turns out, Kubernetes — the orchestrator born from the sudden rise of containerized infrastructure launched by Docker — has very quickly become the engine of workload deployment on cloud platforms. (The word “hybrid” doesn’t have much use anymore in this context, like the prefix in “horseless carriage.”)  Although it was born at Google from a project based on Google technologies (and although Google arguably tried to maintain a rein on its marketing message for some time), no single company owns Kubernetes. Its creators now work for Microsoft (Brendan Burns) and VMware (Joe Beda and Craig McLuckie), and it is overseen by the Cloud Native Computing Foundation, an agency of the Linux Foundation.  VMware has already announced its intention to steer the development of vSphere — the virtual underpinnings of a huge chunk of enterprise infrastructure — toward Kubernetes.

Red Hat had a stake in this game early on, long before competitors decided to shift their strategies towards Kubernetes. From this entrenched position, it hopes to be able to solidify that stake by appealing to the people who brought Red Hat into the enterprise in the first place: middle managers,  IT administrators, and so-called “transformation team leaders.” At the heart of Red Hat’s message is affirmation more than automation — the idea that tasks and workloads in an automated deployment system can be owned and appreciated, rather than outsourced to a machine.

“No one likes to hear that what they do is bad,” explained Shafer. “Humans tend to attach their identities to their tasks. And as a consequence of this, a lot of times, what you’re dealing with in organizations is actually, people hearing that what they’re about to do is erase their identity.”

This is how resistance builds up in the organization — specifically, resistance to change and a gumming up of the activities chain of organizations for which middle managers may inevitably be blamed. Notice how deftly Red Hat has transformed (to borrow its actual term for the office from which this message emanates) the topic of infrastructure deployment from an automation issue to a cultural issue.

“To me, this is like systems thinking,” Shafer continued. “You can try to push things harder, or you can try to remove the resistance to change. The more that you can make people see themselves as heroes in the new version of this story, then the more receptive they are to building these bridges to a new kind of work.”

War stories

Red Hat’s intent with gatherings such as its OpenShift Commons, which was being held this year on the Inspiration Hornblower yacht, has been to encourage OpenShift adopters in the enterprise to share and discuss their deployment experiences — what Red Hat Open Innovation Labs director Kris Pennella calls “the good, the bad, and the ugly” — so that the open-source process can extend beyond software development and into infrastructure deployment. It’s another way to solidify Red Hat’s stake in this rapidly emerging Kubernetes market, by encouraging implementers from ING, ExxonMobil, and Broadcom, among others, to get to know one another and collaborate.

In a world whose infrastructural underpinnings, in whatever form they may take, may all be provided by the same engine, it becomes necessary for each participating vendor in this market to forge a community around its brand and its ethos. This is where Red Hat hopes it already has a four-year head start on VMware.

“It’s pretty easy to identify how you’re a hero as a developer,” explained Jabe Bloom, who works in Red Hat’s Global Transformation office. “It’s pretty easy to figure out who can be a hero as a CEO. You go into any airport, and there’s like a thousand books on how to be a hero as a CEO. Middle management is completely devoid of stories about heroes, as middle managers. If you go into most organizations and ask what do middle managers do, you get, ‘I don’t know, they get in the way.’ The number one indicator that your transformation has been successful is the increase in peer-wise communications in middle management — someplace where they can practice talking to each other and practice telling their stories.”

One organization that claims to have had some success in recent months getting management on-board with changes to its digital infrastructure, is ExxonMobil. Audrey Reznik is a member of that company’s Computational Data Scientists Group, which applies simulations and machine learning approaches to the studies of geology and mineralogy. As Reznik told attendees, as late as 2017, it was a challenge for her team to produce and present proofs of concept (PoC) in a form that users and colleagues could actually use. White papers and published reports make for nice references on one’s curriculum vitae. Yet “who wants to sit there and install a lot of software,” remarked Reznik, “when you just want to see if the problem that you have can actually be solved. Or, is it a problem that you should even be looking at?”

Jupyter Notebooks gave team members environments in which to paste code, stage quick tests, and show results. But the patience levels of her intended users still have their limits. An ExxonMobil colleague suggested during lunch that she create a container for her PoC code, rather than a short-lived notebook, using OpenShift.

“So instead of worrying about giving people local admin access, or worrying about the latest source code, or even worrying about some of the dependencies that you have,” she continued, “you can contain them into an atomic unit that you can go ahead and deploy. And we sat back and said, if this works for us, we would have something that is both reproducible and interactive. It would save us the trouble of flying to our colleagues in Calgary, and setting up their laptops just so they could see the latest solution.”

Under the old system of demonstration, last year, Reznik’s team managed to produce two minimum viable products (MVP) — PoCs built with just enough features to cross the border into usable, software-based experiments and simulations. After implementing OpenShift, she told attendees, her team had produced over 70 MVPs since last January.

The move to OpenShift ended up prompting the Computational Data Science team to adopt a DevOps-like approach to software development, utilizing Git repositories as a means for implementing staging and version control. “For some of the scientists,” she said, “this was a new thing.”

OpenShift, coupled with the tools built around Kubernetes, presented these data scientists with the means to deploy concepts built-in Jupyter notebooks, develop and evolve them through Git repositories, and then deploy them to the cloud in such a way that colleagues and users could utilize them by typing the URL into their browsers. In devising a kind of itinerary for building these PoCs in a controlled manner, the team ended up adopting Agile software development methodologies. “For us, this was actually groundbreaking,” said Reznik. “It may not be for you, but for us, this was huge.”

So it was that, over lunch one day, some of the most sophisticated software infrastructure ever devised was put to use in delivering something as seemingly simple as a proof-of-concept. A few dozen of these little breakthroughs a day, and after a time, IBM may find that OpenShift has delivered a return on its Red Hat investment.

It’s the kind of story best told to your shipmates.

Disclosure: Scott Fulton’s appearance at OpenShift Commons and KubeCon 2019 was made possible by the main conference organizer, the Cloud Native Computing Foundation.

Article source: https://www.zdnet.com/article/ibm-paid-34b-for-this-where-red-hat-is-taking-openshift/#ftag=RSSbaffb68