What is distributed and clustered?What is the difference?

Go to: https://blog.csdn.net/jiangyu1013/article/details/80417961

A few days ago, I found a giant cow's artificial intelligence learning website. It is easy to understand, funny and humorous. I can't help but share it with you.Click to jump to the tutorial.

PS: This article is a note, only part of the text is original, and quite a lot of content is just collected, sorted, refined, and summarized by others.

Not marked as original out of respect and thanks to the original authors of some of the content...

But if you want to reprint, please bring my home address. After all, these contents are not ready-made. ~~ ~

I. Explain in vernacular, you can understand it in half a minute, and use examples in life to illustrate:

The small restaurant used to have only one chef, who cut and washed vegetables, prepared and cooked all the vegetables.

Later, there were too many guests, one chef in the kitchen was too busy, and another chef was hired. Both chefs could cook the same dishes. The relationship between the two chefs was a cluster.

In order to let the chef concentrate on cooking and make the dishes to the extreme, we hired a side chef to be responsible for cutting, preparing, and preparing the vegetables... The relationship between the chef and the side chef is distributed.

A side chef is too busy, and another side chef is hired. The relationship between the two side chefs is a cluster.

A side chef has asked for leave for some reason, but the rest of the side chefs still do whatever they need to do. It's just that the tasks of the side chefs who didn't ask for leave are evenly increased, but their tasks and responsibilities remain the same.This is the cluster.

Cluster: Multiple people doing the same thing together.

Distributed: Multiple people working together to do different things.

2. Diagram:

The above content is reproduced from: https://www.zhihu.com/question/20004877

The following is personal curation:

Third, distinguish the connection (in fact, the above content should have already made you understand the two)

1) I remember there is such a sentence in a book about tcp/ip: Distributed refers to a system in which multiple systems cooperate to complete a specific task.

Distributed is to solve the problem of centralized management. It is too slow to superimpose all tasks to a node for processing.

So divide a big problem into multiple small problems, solve them separately, and finally work together.The main work of distributed is to decompose tasks and disassemble functions.

2) The main usage scenario of a cluster is to share the pressure of requests, that is, to deploy the same application on several servers to share client requests.

When the pressure increases further, mysql may not be able to face a lot of write pressure in the part that needs to be stored.Because after mysql is formed into a cluster, the main writing pressure is still on the master machine, and other slave machines cannot share the writing pressure, so at this time, distribution is also introduced.

The main application scenario of distributed is that a single machine can no longer meet the performance requirements, and multiple nodes must be integrated, and the nodes are related and interacted.It is equivalent to when writing mysql, each node stores part of the data, which is the origin of distributed storage.Store some unstructured data: static files, pictures, pdfs, small videos... These are the origins of distributed file systems.

3) The cluster is mainly to simply add machines to solve the problem, without any decomposition of the problem itself;

Distributed processing must include task decomposition and answer merging.A subtask node in the distribution may be replaced by a cluster; any node in the cluster is doing a complete task.

Cluster and distributed are composed of multiple nodes, but the communication coordination between clusters is basically not required; and the communication coordination of distributed nodes is essential.

Split a system into different subsystems and deploy them on different servers (this is called distributed),

Then deploy multiple identical subsystems on different servers (this is called clustering). The same subsystem deployed on different servers should be load balanced.

Distributed: A business is split into multiple sub-businesses and deployed on multiple servers.

Cluster: The same service is deployed on multiple servers.

A few days ago, I found a giant cow's artificial intelligence learning website. It is easy to understand, funny and humorous. I can't help but share it with you.Click to jump to the tutorial.

Reference: https://www.zhihu.com/question/20004877