Is Kubernetes a perfect defense against cloud lock-in, or a needlessly complex solution?
Oh, to be sure, Kubernetes are very popular, of course you need one, maybe three! However, if you keep thinking: Why Kubernetes? Then you would expect a lot of comments.
One of the main reasons IT professionals cite Kubernetes is to reduce locking by ensuring portability between clouds. In theory, this is better than in practice. And, as Johnston says, the people who told him to choose Kubernetes for cloud portability told him they had no plans to migrate.
So why use Kubernetes?
Containerized lock!
A lot of people find themselves on the Kubernetes trend because it’s popular. (” Developers and architects use it because technology is a fashion industry, and Kubernetes is popular, “Orion Edwards says.) Still, James Thomason argues that while developers may see Kubernetes as a “way of acting like Google”… In fact, with the exception of 0.001 percent, this is unnecessary.”
Although this may be an exaggeration, Thomason has a point. As an industry, we do tend to take shiny new things far beyond their intended use.
According to Johnston, many ctos embrace Kubernetes “usually because they have to.” Either it was inherited or it was the next big thing in their eyes (lots of developers are hiring)
Why regret? Because Kubernetes introduces a level of complexity that the low-level Docker container, the cloud portability tool they need most, does not. Or a simple shell script. In fact, as Johnston puts it, Kubernetes ended up “complicating, in many different ways, what they had been doing for years.”
People respond to Johnston’s “Why Use Kubernetes? The main answer is to avoid locking. As Dan Selman sees it, “It’s not always a rational fear, but it is a fear.” “There is a legitimate fear of lock-in,” said Analyst Lawrence Hecht. Even if you’re not going to use an exit strategy, it’s reasonable to want to have one.”
Do you want cloud portability to minimize locking? You can have it. But you probably don’t need Kubernetes to get there.
From Johnston’s point of view, attempts to evade locking should not “automatically mean Kubernetes. We installed the overall components on the virtual server to achieve portability. I think we have less portability with Kubernetes now.”
More Kubernetes, less portability? How does this work? “There are ways to reduce your application’s dependence on these things by subtly leveraging some of the Kubernetes apis, but in general, using bare Kubernetes doesn’t give you cloud portability for free,” Neal Gompa said.
Kubernetes, behind the scenes,
Even if Kubernetes does not remove real-world locks, it is still valuable for other reasons. On the one hand, if developers are based on Kubernetes, they will gain valuable skills that can be transferred between different employers, regardless of what cloud they are using.
Furthermore, Kubernetes is a good way for an enterprise to capture the level of abstraction of the infrastructure, even if it does not eliminate locking, to help when moving between services. After all, most businesses have a reason to choose specific clouds over basic computing and storage.
As an industry, we have a tendency to focus on technology, even as vendors are eliminating that focus. James Urquhart, for example, insists that Kubernetes will win, but it’s impossible to get every developer to install and use it. Instead, he suggests, “[Kubernetes] should be completely hidden under important abstractions.
In other words, developers may end up using Kubernetes behind the scenes, drowning in server-free products and the like. But most people don’t have to dig into the Kubernetes API.
Does this mean Kubernetes will fail? No, Kubernetes will win when it reverts to its “invisible pipe” state.
The above content is some of my own feelings, share out welcome correction, incidentally beg a wave of attention, have ideas of partners can comment or private letter I oh ~
Translated by Zhang Chun
Source:
The new titanium cloud service