How to maximize the concurrency capability of the PHP server within limited resources

Assume that the application app uses 5W rmb to purchase from suppliers. On the day of registration, a large number of candidates flood in, and the number of concurrent requests soars to 30W, causing the system to shut down and deny service, causing the candidates to suffer Unable to register, can 5W rmb support 30W concurrency?

But for us, we might as well raise the problem to another angle: "How to maximize the server's concurrency capability within limited resources." Suppose you are a technical leader, how would you design and structure a project with a large amount of concurrency?

First of all, we can give a general idea of ​​​​this project. From the above description, it is not difficult to see that the bottleneck of this project is "concurrent writing" rather than "reading", so from the resource allocation point of view, we can " "Write" tilt, here I write all the data in Redis. In addition, we also need to try our best to migrate MySQL read operations to Redis. The work done by MySQL is more inclined to some conventional non-concurrent read and write operations.

Server

When a user request comes, the load balancer loads it to each server

This is a stress test data from symfony, using the configuration of 1 CPU, 4 GB and PHP 7.

The data in the above picture comes from the swoole official website. After adding the execution of our actual business logic, we can find that when we use the resident memory startup method , 3 lower-end servers can solve the above problem that requires 16 servers.

Database

In fact, many people will understand that the bottleneck of many current Internet projects is more concentrated in database I/O. In this area, there is no particularly big gap between the various languages. Including the widely criticized PHP-FPM startup method, you can also use swoole and other methods to replace it. Therefore, in this project, we will focus more on the database, and we can try to use Redis to solve it. Of course, in the specific code, we also need to prepare a certain number of data connection pools in advance. In addition, it is also considered that although MongoDB's writing speed is much faster than MySQL under the same configuration, it still has obvious shortcomings compared to Redis.

Register and log in

Registration and login should actually be divided into two parts, which correspond to "write" and "read" respectively. In the case of high concurrent reading and writing, using MySQL directly will explode as you expected. Therefore, we can cache user data into Redis during the process of building the entire project. "Writing" problem: When the number of users is unclear and the amount of concurrency is large, I prefer not to store user data directly in the database. We can design a switch or threshold to set the user's warehousing method. When the concurrency is large, users can be asynchronously warehousing through MQ, but in normal times, users can be warehousing normally.

Submit form

Because this project is not a common flash sale and requires immediate notification, the design of our project greatly reduces the difficulty. The function of submitting the form is similar to registration. We can completely load the data into the database asynchronously and then review it in the background.

Summary

I won’t go into details about whether other CDN and MySQL require master-slave, it depends on the actual situation. Theoretically, if you use PHP-FPM, it will cost about 19,000 yuan/month to solve this problem of the project. When using swoole, it will cost about 4,500 yuan/month. I am not advocating swoole here. What I want to explain is that when When we face large concurrent projects, especially when the business logic is relatively complex, we can better solve the problem by using resident memory, and this has nothing to do with language. Finally, it should be noted that the above is only a theoretical stage, and actual data needs further testing. The material of the article comes from the Internet. If there is any incorrect writing, please point it out.

The above is the detailed content of How to maximize the concurrency capability of the PHP server within limited resources. For more information, please follow other related articles on the PHP Chinese website!