Shulou(Shulou.com)06/01 Report--

This article shows you what are the high availability and load balancing architecture design patterns commonly used in big data, which are concise and easy to understand, which can definitely brighten your eyes. I hope you can get something through the detailed introduction of this article.

Here is a summary of common high availability and load balancing architecture patterns.

1. Client switching.

The client configures multiple server addresses. If a server becomes unreachable or uses an exception, the client switches to another server.

Advantages: simple, efficient, and can test service availability at the business level

Disadvantages: need to be configured in advance.

Mysql connect does this for load balance and failed over. Multiple servers can be configured on a JDBC connection.

# Connection URL for a server failover setup: jdbc:mysql//primaryhost,secondaryhost1,secondaryhost2/test2. DNS mapping

The DNS service returns different IP to achieve the effect of switching and load balancing. Because DNS caches, there will be more problems in actual use. .

Advantages: transparent to users

Disadvantages: due to DNS cache problems, switching may not be timely

3. Service registration discovery

The registration server (zookeeper) is used to manage the currently available service offerings. Dubbo is implemented in this way.

Advantages: complex load balancing strategies can be used to implement service routing.

Disadvantages: complex implementation

4. Keepalive+ floating IP

If there is a problem with the machine in use, the external network Ip can be assigned to another host according to the situation. This makes the IP switch transparent to external visitors. However, to implement floating IP, support for the underlying network. It cannot be used in current CVM.

Advantages: efficient and transparent to the client

Disadvantages: there is a requirement for the host network and cannot be used in the CVM.

5 proxy server (nginx, lVS,Haproxy)

The proxy server detects whether the upstream server is available. And the request from the front end will be routed according to a certain policy. In order to avoid the failure of the proxy server, the system cannot be accessed. Therefore, it is generally used in combination with keepalive+vip.

Advantages: support a large number of servers and configure complex routing rules

Disadvantages: the proxy server has a single point of problem and fault detection is not timely.

To sum up, the main differences of the above models are:

1. Server information who maintains, client, proxy server or registration service

two。 The means of fault detection. The general detection method can only check whether the port can be connected, and if the server zombie occurs, it can not be found in time.

3. Server utilization.

What are the high availability and load balancing architecture design patterns commonly used in big data? have you learned any knowledge or skills? If you want to learn more skills or enrich your knowledge reserve, you are welcome to follow the industry information channel.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.