当前位置：网站首页>Differences and introduction of cluster, distributed and microservice

Server cluster is a kind of server system , It works through a loosely integrated set of server software and / Or the hardware is connected together to complete the computing work closely . In a sense , They can be seen as a server .

A single server in a cluster system is usually called a node , Usually connected through a local area network , But there are other possible ways to connect . Cluster servers are usually used to improve the computing speed and performance of a single server / Or reliability . In general, cluster

The server is better than a single server , For example, the performance price ratio of workstation or super server is much higher . A cluster is a group of independent servers , Through network connection, combine into a group to complete a task together .

Speak straight and white , A cluster is a set of independent servers , Through the high-speed network to form a server system , Each cluster node is a separate server running its own process . For Internet users , The back end of a website is a single system , Work together to provide system resources to users , system service .

Two 、 Why use clustering ？

1、 Characteristics of cluster

- High performance performance

Some require a lot of computing power, such as weather forecasting , Nuclear tests, etc . It's not something a few servers can do . It takes thousands of them to complete the work .

- Price effectiveness

Usually a system cluster architecture , Only a few or dozens of server hosts are needed , Compared with mobile, millions of dedicated super servers are more cost-effective .

- Scalability

When the server load pressure increases , The system can be extended to meet the requirements , And it doesn't reduce the quality of service .

- High availability

Despite some hardware and software failures , The service of the whole system must be 7*24 Running in hours .

2、 Advantages of cluster

- transparency

If some servers are down, the business will not be affected , Generally, the coupling degree is not so high , Dependency is not that high . such as NFS If the server goes down, the rest can't be mounted , It's too dependent .

- High performance

More visits , It's easy to expand .

- manageability

The whole system can be physically large , But it's easy to manage .

- Programmability

On a cluster system , Easy to develop applications , The portal will ask for this .

3、 Cluster classification and the characteristics of different classifications

Computer cluster architecture is generally divided into the following categories according to its function and structure ：

- Load balancing cluster （Load balancing clusters） abbreviation LBC

- High availability clusters （High-availability clusters） abbreviation HAC

- High performance computing cluster （High-perfomance clusters） abbreviation HPC

- Grid computing （Gridcomputing）

In terms of cluster classification , There are generally three on the Internet , Load balancing and high availability cluster architecture is commonly used in our Internet industry .

3、 ... and 、 Cluster classification introduction

1、 Load balancing cluster

Load balancing cluster provides more practical for enterprises , More cost effective system architecture solutions . The load balancing cluster allocates the load pressure of requests from many customers to the computer cluster as evenly as possible .

Customer request load usually includes application level processing load and network traffic load . Such a system is ideal for serving a large number of users using the same set of applications . Each node can bear the load pressure of access request , And it can realize the dynamic allocation of access requests among nodes , To achieve load balancing .

Load balancing runtime , Generally, one or more front-end load balancers distribute client access requests to a group of servers on the back end , So as to achieve high performance and high availability of the whole system . This cluster is sometimes called a server cluster .

Generally, high availability cluster and load balancing cluster will use similar technology , Or it has the characteristics of high availability and load balancing at the same time .

The role of load balancing cluster ：

a） Share access traffic （ Load balancing ）

b） Maintain business continuity （ High availability ）

2、 High availability clusters

Generally refers to when any node in the cluster fails , All tasks on the node are automatically transferred to other normal nodes , And this process does not affect the operation of the entire cluster , Does not affect the provision of business .

Similarly, there are two or more nodes running in the cluster , When a master node fails , So what else can we do The node of the node will take over the master node , Continue the task above . The slave node can take over the resources of the master node （IP Address , Architecture, identity, etc ）, At this time, the user will not find that the service object is transferred from the master node to the slave node .

The role of high availability clusters ： When one machine goes down and another takes over . More commonly used high availability cluster open source software is ：keepalive,heardbeat.

3、 High performance computing cluster

High performance computing cluster adopts the method of assigning computing tasks to different computing nodes in the cluster to improve the computing power , So it is mainly used in the field of scientific computing .

More popular HPC use Linux Operating system and other free software to do parallel computing . This cluster configuration is often referred to as Beowulf colony . Such clusters usually run specific programs to play HPCcluster The ability to parallel . This type of program usually uses a specific runtime , For example, designed for scientific computing MPI library .HPC Cluster is especially suitable for computing jobs where a large amount of data communication occurs between computing nodes , For example, the intermediate result of one node or the situation that affects the calculation result of other nodes .

Four 、 The difference between cluster and distributed

Cluster is a physical form , Distributed is a way of working

Distributed ： One business is split into several sub businesses , Each sub service is deployed on different servers
colony ： The same business , Deployed on multiple servers

1） Distributed refers to the distribution of different sub businesses in different places . Clustering refers to clustering several servers together , Achieve the same business .

2） Every node in the distribution , Can do cluster . Clusters are not necessarily distributed .

give an example ： For example, sina.com , More people visited , He can make a group , Put a response server in front , The following servers complete the same business , If there is a business visit , Response server to see which server load is not very heavy , Which one will be given to complete .

3） And distributed , Understand... In a narrow sense , It's similar to clustering , But it's more loosely organized , It's not like clustering , There is an organization , A server is down , Other servers can be on top .

4） Every distributed node , All complete different sub businesses , A node is broken , Then this business is not accessible .

In short , Distributed is to shorten the execution time of a single task to improve efficiency , Cluster is to improve the efficiency by increasing the number of tasks per unit time .

A good design should be a combination of distributed and cluster , First distributed and then clustered , The specific implementation is to split the business into many sub businesses , Then cluster deployment is carried out for each sub business , In this way, if there is a problem with each sub business , The whole system will not be affected at all .

1、 Distributed diagram

Divide a large system into multiple business modules , Business modules are deployed on different machines , Each business module interacts with each other through the interface . The way to distinguish between distributed systems is based on different machines and different services .

above ：service A、B、C、D Business components , adopt API Geteway Business access .

notes ： Distributed business needs to be well managed

2、 Illustration of cluster mode

Cluster mode is that different servers deploy the same set of services for external access , Load balancing of services . The way to distinguish clusters is based on whether the services of multiple servers deployed are the same .

notes ： Cluster mode needs to be done well session share , Make sure that in the process of switching between different servers, it will not be caused by not getting session And discontinue the service .

General configuration Nginx* Implementation of load container based on XML ： Static resource cache 、Session Sharing can come with ,Nginx Support 5000 Concurrency .

5、 ... and 、 The difference between distributed and micro Services

in addition , Another concept is similar to distributed computing , That's Micro service .

Microservice is an architectural style , A large complex software application consists of one or more microservices . Each microservice in the system can be deployed independently , The microservices are loosely coupled . Each microservice only focuses on accomplishing one task and accomplishing it well . In all cases , Each task represents a small business capability .

1、 Is distributed a micro service ？

Distributed must be microservices , Microservices are not necessarily distributed

The definition of distributed ： Split a service into several sub Services , On different servers . Microservices can be placed on the same server , It can also be placed on different servers .

2、 Microservice architecture

Microservices are designed to avoid the upgrade and BUG Affect the existing system business . The subtle difference between microservices and distribution is , Microservice applications are not necessarily distributed on multiple servers , It can also be the same server .