There was an interesting conversation recently between Ginni Rometty, CEO of IBM, and Jim Whitehurst, CEO of Red Hat, which IBM recently made a $34 billion bid to acquire. Given Red Hat’s history selling Linux and its recent successful foray into cloud-native applications with OpenShift, it was not surprising to hear Whitehurst say the following:
“Whether it’s Linux, containers or Kubernetes (k8s) … that’s where the vast majority of innovation is happening. That’s what autonomous driving is going to be built on. That’s what almost any net new application will be built on. You need to have a robust, high-performing, but also relatively self-managing, self-healing infrastructure at the edge. That’s what Linux, containers or k8s can give you.”
It’s not surprising to see Red Hat, a focused vendor of modern data center software tools, pitch containers or Kubernetes. What is surprising is that the CEO of IBM, a broad-based technology company, should reinforce that containers are key to modern enterprise IT and essential to moving the needle on IBM’s massive $75 billion in revenue.
“You have to take mission-critical processes and applications and put them in a container in a k8s, micro-services world,” Rometty said in response to Whitehurst. “That is what is coming ahead of us.”
I’ve been arguing in my column here recently that Kubernetes is the next big enterprise platform. The reason is that enterprises are under enormous pressure to increase agility as they face competitive pressures. Kubernetes, along with corresponding cloud-native practices like micro-services, DevOps, and continuous integration and development, has emerged as a platform that enables teams to build apps more quickly and run them at scale, empowering developers.
For instance, the lead application architect for T-Mobile was recently asked what was different about how T-Mobile built and ran applications on Kubernetes versus before. His answer was enlightening: “What drove us to containers in the first place was simply that in order to get code from the development environment into production took up to six months and 72 steps,” he said. “[Now] we’ve had customers that can reach out to us, give us their requirements [and] we onboard them that same day. They can push an application to production that same day.”
So the power of Kubernetes to streamline development and operations is compelling. But do Kubernetes’ agility benefits extend across enough applications to make an impact on the organization overall?
Put another way, a typical enterprise might have 10,000 applications powering its various lines of business. Does Kubernetes allow an organization to increase the agility of a sufficient proportion of those apps to make a dent that is noticeable at the board or shareholder level? This is what Rometty needs to have happen if her company’s $34 billion bet is to pay off.
Unfortunately, I believe the answer is no if Kubernetes can’t move beyond applications at the periphery of an organization to the core. This is exactly where I feel Kubernetes is today: on the cusp of greatness, as Rometty hopes, but not quite there yet because it alone does not solve the non-negotiable, hard business requirements that come with most enterprise applications.
While it’s true that Kubernetes usage is growing rapidly, the applications are typically web-style applications (websites, mobile apps) or relatively simple line-of-business applications. And these apps make up a tiny portion of the overall enterprise portfolio, and they don’t tend to have complex business requirements such as strict data security, SLAs, disaster recovery, compliance and governance, and so on.
What this means is that the apps being containerized today are simpler, thus the gains from improving their agility are smaller, and the money to be made supporting them by companies like IBM is limited.
If Kubernetes is only used for less complex applications, it will exist only as a niche platform. The history of application platforms bears this out. Microsoft won the PC wars in part because of its strong distribution arrangements, its enterprise support, and by having more third-party apps than Apple. Linux, AWS, Salesforce and VMware are all examples of successful platforms that gained similar widespread adoption through their strong ecosystems, add-on solutions and ability to go beyond simple applications.
Kubernetes is well-positioned to be the enterprise application platform of the next decade and fulfill Rometty’s prediction, but only if enterprises can easily run mission-critical applications on the platform with the strict, non-negotiable business requirements I described above. Enterprises will need to add solutions targeted at these problems to the core Kubernetes platform to get the full value of containerization.
>> Source Link