foreword：

I believe everyone knows，Talking about the design scheme of the high-quality framework of the billion-dollar traffic website，Two key points come to mind，That is high availability and high concurrency。And talk about how familiar you are with high concurrency and high availability，Jingdong's bosses still have a certain say，And as a senior engineer at JD.com, he is even more powerful。

What I share today is written by an engineer with ten years of development experience in JD.com：“Core technology of 100 million-level traffic website architecture”。

note author：Jingdong senior engineer：Big fly

Notes Features：well organized，with graphic，The content is innovative and esoteric but easy to understand。

The notes are roughly divided into four parts∶Overview、High Availability Principle、High Concurrency Principle、Case combat。load balancer、Limiting、downgrade、isolate、timeout and retries、rollback mechanism、Stress testing and planning、cache、pooling、Asynchronous、Expansion、The core technology of the architecture of the billion-level traffic website is introduced in detail in many aspects such as queues，I believe that after reading it, you can quickly apply it in practice。

Don't talk too much, just take notes：

Special Note：Because the notes are too detailed，Due to limited space，The contents of this document can only be displayed in the form of pictures below.，if you need“Full HDpdfVersion”，You can directly private message me in the background to get keywords「Architecture」Get the free way。

the first1 part Overview：

1. Some principles of transactional system design

2. High Concurrency Principle

(jd website development)

• no status
• split
• Servicing
• message queue

• data heterogeneity
• cache silver bullet
• concurrency

3. High Availability Principle：

• downgrade
• Limiting

• cut flow
• rollback

4.business design principles：

• Heavy-duty design
• Idempotent Design
• Process can be defined
• State and State Machine
• Back-end system operation feedback
• Back-end system approval
• Documentation and Notes
• backup

the first2 part High availability：

1.Load Balancing and Reverse Proxy：

• upstreamconfigure
• load balancing algorithm
• retry on failure
• health examination

• Other configuration
• Long connection
• H T T PReverse proxy example
• H T T PDynamic Load Balancing
• NginxLayer 4 load balancing

2. isolation：

• thread isolation
• process isolation
• Cluster isolation
• Computer room isolation
• Read and write isolation