Job interviews are always a second kill? After reading the seckill system notes secretly stored by JD T8, I have given my knees

Whether it's an interview or at work , We are all prone to problems ：

• How to achieve capacity expansion under extreme process pressure ？

• When the system crashes , After the service is blown , How to ensure the normal operation of data ？

• When the system is new , How to ensure the stability of the system ？

• How to realize the high availability architecture design required by each level ？

• How to implement a system that can support high concurrency , Highly available system ？······

Understand the underlying principles , In practice, he is a master . For example, high concurrency means that a large number of visitors visit the target system in a relatively short time , System load saturation or overload downtime . High concurrency applications , We should all have used or seen , For example, Tmall. 、 JD.COM 、 A lot of spelling 、 Amazon's second kill rush and 12306 Ticket grabbing .

When we are experiencing the application , You may not think of the difficulty of implementing the technology behind this highly concurrent system . These problems exist in high concurrency systems , High concurrency high concurrency read and write 、 Sudden access peak 、 Immediacy of feedback results .

However, from 0 To 1 Achieve a high performance , High concurrency , Second kill system with high availability , There must be a lot of things to pay attention to , Share a learning note document of seckill system architecture for free here ！

First look at the catalogue

One 、 What are the key points in the architecture design of spike system ？

Second kill mainly solves two problems , One is concurrent reading , One is concurrent write . The core optimization concept of concurrent reading is to minimize the number of users to the server “ read ” data , Or let them read less data ; The same principle applies to concurrent writes , It requires us to separate a database at the database level , Do something special .

Two 、 What should be paid attention to when designing the second kill system 5 There are two principles of architecture

Seckill system is essentially a system that satisfies large concurrency 、 High performance and high availability distributed system . This chapter mainly talks about how to meet the requirements of a well structured distributed system , To achieve the ultimate performance improvement for the second kill business .

3、 ... and 、 How to do a good job of dynamic and static separation ？ What are the options ？

Key points of this chapter ： The separation of static and dynamic data . I wonder if you've heard of this solution before ？ Whether you've heard of it or not , I suggest you stop and think about the value of the separation of motion and static . If your system has not yet started to apply the dynamic static separation scheme , Then you can also think about why there is no , I didn't think of it before , Or is there no need for business volume ？

Four 、 This principle ： Deal with the problems of the system “ Hot data ”

5、 ... and 、 How to cut the flow peak ？

If you have seen the flow monitoring diagram of seckill system , You'll find it's a straight line , It's a very straight line at the beginning of seckill , This is because the seckill request is highly focused on a specific point in time . thus , It will lead to a very high traffic peak , It consumes resources instantaneously

6、 ... and 、 What are the factors that affect performance ？ How to improve the performance of the system ？

This chapter combines the second kill scene , Focus on introducing some optimization skills of the server .

7、 ... and 、 Seckill system “ Reduce inventory ” Core logic of design

If you want to design a second kill system , Then I think your boss will tell you first ： Never oversold , That's the big premise

8、 ... and 、 How to design the bottom plan ?