Large site technical architecture, website architecture evolved under the course

1.2.6 Using reverse proxy and CDN acceleration response website

To solve the problem is poor regional network

CDN and the basic principles are the reverse proxy cache,

CDN deployed in the network room network provider, users request Web site service, you can get data from the network provider from their recent room

Reverse proxy are deployed in the central office site, a user requests a web service, when accessed first reverse proxy server, if the reverse proxy server caches the resource requested by the user, it will be directly returned to the user.

The purpose of using a reverse proxy server and CDN are returned as quickly as possible to the user:

1, to speed up user access speed

2, to reduce the back-end server load pressure

1.2.7 using the distributed file system and distributed database systems

Distributed data used only in very large-scale single-table data when, less than a last resort, when the site more commonly used database splitting means business sub-libraries, databases of different services deployed on different physical servers.

1.2.8 use NoSQL and search engines

Select the scalability of NoSQL distributed and good search engine server selection.

In addition the application server to access a variety of databases via a unified data access module, reducing application management trouble many data sources.

1.2.9 Business Split

Means and rule by the entire sub-site business into different product lines, such as Taobao will be home, shop, order, buyers, sellers, etc. split into different product lines, divided into different business groups responsible.

The application service splitter, and deployed on different servers, the distribution data through the message queue. Or forms a complete system by accessing associated with a data storage system.

Stateless when the service server, the calling context does not store data, a transaction completed through multiple application servers, such as ordering systems and inventory systems, order success is the success of data into single message queue, the message queue to be guaranteed scalable, high availability, high-performance, inventory system on the server Redis Save inventory, inventory system subscription to a commodity single topic, once received orders will start logical inventory reduction, if inventory system receives a message linked , do not confirm the message, the message still remains in the message queue, the message queue if the message is processed normally, manually confirm this news consumed.

 1.2.10 Distributed Services

Some public service modules, the number of calls is very large, you can take away a module, independent deployment alone.

Business system is only responsible for its own business logic, data, public services need to call, went public business service call to complete.

 

 Since a large site architecture to solve the massive data management and high concurrent transactions, then you can apply these solutions to business outside of the site itself up.

At present, many large sites are beginning to build their own cloud computing platform, computing resources as a basis for the sale, small sites do not need to be concerned about the technical architecture issue, simply pay-as-you can make the site as the business grows gradually gained more large storage space and more computing resources.

 

Guess you like

Origin www.cnblogs.com/wozixiaoyao/p/11421025.html