The first two installments of the “Kubernetes History” series “Is Kubernetes a Gift from Google?” And “Kubernetes’ Achilles heel” respectively show Kubernetes’ important value and significance in many aspects from the imaginary future 50 years later and the real present. Today’s story is from a more distant future. Looking back, will Kubernetes go the same way as openstack?
On Thursday, Link woke up early and took a flight to Guizhou.
Link spends most of his time in Yunnan, but occasionally travels abroad. Because Thursday is a working day, I will go to Guizhou once.
There are not many people in Guizhou. It is true that the temperature here is a bit cooler than in the south.
In 2017, 100 years ago, companies chose locations for their data centers because they were cheap, cold and far from people.
However, Link is not a social butterfly, preferring to spend more time alone and walking with the animals in the rainforest.
So, data center engineer is a good fit for him, after all, this kind of functional job has disappeared after the fifth information revolution.
Along came more soul chaperones, dating experts, sociologists, artists, fashion experts.
That’s understandable, given that compared to a century ago, people had to work five days a week to make a decent living.
Now, just one day a week can make a good living.
People have more days to socialize, to look at art, to follow trends.
When they arrived at the office, only a few robots walked back and forth, idly saying a gruff “Hello, Link” and continuing to walk back and forth.
His 18th-floor office had a floor-to-ceiling window, out which he occasionally looked, unchanging, at the spotless road, the flowing man-made river and the mountains beyond.
Put the Box on the workbench (Box looks like a rectangular glass block, about 5 inches) and wake up the workbench, which is a 40-inch touch-screen.
The console contains several kanban panels: physical topology, resource management, application management, etc.
Each image can be scaled.
Each layer you zoom in on reveals one more detail of the system, while some details are hidden when you zoom out.
Most of the time, Link sees the big picture and only cares about the details when something goes wrong.
This top-down model seems natural.
Link, however, has read history and knows that the top-down model didn’t become popular until the last century.
The year 2050 is the age of the knowledge explosion, but in 2000, the characteristics of the information explosion are already evident.
There is a time lag from information to knowledge, and it is exactly 2050 that heralds the age of knowledge explosion.
The consequence of the knowledge explosion is that no one can learn a subject from beginning to end. As Lao Tzu said, “Knowledge is long, life is short. With the end with the end, almost already!”
So top-down learning prevailed, in which everyone first showed interest in a subject and mastered the necessary elements of knowledge,
Then we will learn the details under the guidance of machine assistants. The machine is the external store of human knowledge.
As long as people acquire the thinking mode of the subject, and then rely on external knowledge reserve, they can deduce a whole knowledge tree.
Link has been obsessed with Lego since childhood, from which he found the fun of living two, two and three.
So they become immune to system complexity and have the ability to catch the autumn in a flash.
The data center is a system full of complexity and uncertainty.
If there is a problem in the design, it may affect the whole body at any time.
The evolution of data center is the sweat and blood history of operation and maintenance engineers in countless enterprises.
Eventually the operations engineers freed their hands, then freed their brains, and they disappeared.
Of course, few people remember this operation history now.
They don’t teach this kind of stuff, and developers don’t talk about it.
Link knows all this from the Rare Library, and only he knows the old terms like OpenStack and Kubernetes.
Today’s data centers are a bit like the lego aircraft carriers he built when he was a kid.
Every time a new application is launched, it is like an aircraft taking off from the deck, and the M-world is made up of countless aircraft carriers.
Every time a new intern comes on board, Link tells him, “Welcome to the crew!”
At two o ‘clock in the afternoon, Link trained W, the new intern.
Link explains the construction of a data center. “Here we are on deck.” Link uses his hand to draw a model that resembles an aircraft carrier.
Then he began to disassemble the aircraft carrier layer by layer. After building his own things well, he disassembled layer by layer, making Link have an illusion of tearing up the beauty for people to see, and like cutting steak, something that looks complicated has an effortless feeling.
“Finally, take a look at our engine.” Carrier engines are the linchpin of an aircraft carrier and the heart of a data center.
The engine is an interactive layer of hardware and software, like an engine that converts heat into kinetic energy.
“The history of this engine is the history of the game between hardware vendors and data center engines.” Link looked a little excited.
“A long time ago, there was an open source product called OpenStack. It wasn’t very successful, but it left a big mark in the history of open source.
Its failure is undoubtedly related to its own division. Each hardware vendor claims to support OpenStack, pitching it to large enterprises that need data centers.
However, a huge thing, the good and bad are mixed, and it takes a lot of manpower to operate and maintain an OpenStack.
According to historical data, it is difficult for OpenStack vendors to make money, and the money they earn is also hard money.
The ones who make money are the ones who train.
“
“As they say, when united, they part; when divided, they part. OpenStack is fragmented, which is a headache for hardware manufacturers and enterprises.
The arrival of Kubernetes was a turning point.
Kubernetes really took off in 2018.
At the time, there was a lot of concern that Kubernetes would become like OpenStack.
So prescient people began to propose a standard.
The proposal was approved by the CNCF Foundation.
So the Kubernetes organization developed a standard for hardware vendors, K1.0.
Of course, whether the standards are good or not is a matter of profit distribution.
“Although Kubernetes later split, but because there is an industry standard.
Things that split are modular and pluggable.
The whole ecology is moving towards stability.
The predecessor of the current engine Lambda is Kubernetes.
“
After the complete
annotation
1. Spiritual companion: a consultant who answers people’s doubts
2.Box: intelligent device, built-in CPU GPU and other components, can be embedded into any device for use
3. Auxiliary machine: Another name for Robot
4. Rare Library: Museum of Cultural Relics, which preserves a large number of historical books
5.M World: Matrix, a world made up of the Internet
Author’s brief introduction
Taoge
DaoCloud Software Engineer
A programmer, a person who is selfish, stupid, timid, lovable, and so on. His dream was to roam freely in the ocean and between heaven and earth. Here’s how he describes himself:
Between the front end, the back end, the product engineers.
Layered abstractions in reality, dancing on the graveyard of bugs.
I often pick seashells off the coast of Github and feel smug.
You can kill dragons with Python, you can kill dragons with Shell.
Often practice programmer’s way, often read Man documents.
Click here for a review of past content:
“Kubernetes development history” series
