What are stand-alone, cluster and distributed?

First, a vivid picture ：

Stand alone structure
I think you are most familiar with the single machine structure , When the business volume of a system is very small, all the code should be put in one project , Then the project is deployed on a server . All services of the whole project are provided by this server . This is the stand-alone structure . that , What are the disadvantages of single machine structure ？ Obvious , After all, the processing power of a single machine is limited , When your business grows to a certain extent , The hardware resources of a single machine simply cannot meet your business growth needs . At this point, the cluster mode appears , So let's look down .
Cluster structure
There are various kinds of forced explanations for cluster mode in the world of programming apes , Some of them are beyond your comprehension , It's actually a very simple thing , Just listen to me .
When the stand-alone processing reaches the bottleneck , You just make a few copies of the single machine , This constitutes a “ colony ”. Each server in the cluster is called one of the cluster “ node ”, All nodes form a cluster . Each node provides the same service , In this way, the processing power of the system can be increased several times （ There are several nodes that are up to so many times ）.
But the question is which node will handle the user's request ？ It is better to be able to let the node with smaller load handle at this moment , This makes the pressure at each node average . To implement this function , You need to add one... Before all the nodes “ Dispatcher ” Role , All requests from the user are given to it first , Then it depends on the current load of all nodes , Decide which node will handle the request . This “ Dispatcher ” There's a cow named —— Load balancing server .
The advantage of cluster structure is that system expansion is very easy . If with the development of your system business , The current system can't support , Then add more nodes to the cluster . however , When your business develops to a certain extent , You will find a problem —— No matter how you add nodes , It seems that the performance improvement effect of the whole cluster is not obvious . Now , You need to use the microservice structure .
Distributed structure
Let's make a summary of the previous knowledge points .
From stand-alone structure to cluster structure , Your code basically doesn't need to be modified , All you have to do is deploy more servers , Just run the same code on each server . however , When you want to evolve from a cluster structure to a microservice structure , The previous code needs to be changed a lot . So for the new system, we suggest , At the beginning of system design, micro service architecture was adopted , In this way, the cost of later operation and maintenance is lower . But if an old system needs to be upgraded to a microservice structure , Then we have to fight the code . therefore , For the old system , Keep cluster mode on earth , Or upgrade to microservice architecture , This requires careful consideration by your architect 、 Weigh the input-output ratio .
OK, Let's start with the so-called distributed structure .
Distributed architecture is a complete system , According to the business function , Split into independent subsystems , In a distributed architecture , Each subsystem is called “ service ”. These subsystems can run independently in web In the container , They pass through RPC Means of communication . for instance , Suppose you need to develop an online mall . According to the idea of microservice , We need to split it into multiple independent services according to functional modules , Such as ： Customer service 、 Product service 、 Order service 、 Background management services 、 Data analysis services and so on . These services are independent projects , Can run independently . If there is a dependency between services , Then through the RPC Way to call .
There are many advantages of this ：
The coupling between systems is greatly reduced , It can be developed independently 、 Independent deployment 、 Independent testing , The boundary between the system and the system is very clear , It's also quite easy to make mistakes , Development efficiency is greatly improved .
The coupling between systems is reduced , So the system is easier to expand . We can extend some services specifically . Let's say that the mall will make a big promotion , Orders may be greatly increased , So we can improve the order system 、 Number of nodes in the product system , And for the back office management system 、 Data analysis system , Keep the number of nodes at the original level .
Services are more reusable . such as , When we use the user system as a separate service , All products of the company can use the system as a user system , No need for redevelopment .

Reprint zizhihu user Big idle man, Chai Maomao