AIY商城技术部分英文翻译
Back End
- Support high concurrency, single server supports 10,000 concurrent
- High availability, support for server clusters
- Order system uses queue architecture to solve traffic peak clipping
- Data security, service security
- Support for large data storage
Front End
- Avoid web requests to refresh pages based on Ajax asynchronous requests
- Based on Vue.js framework, single page web application, jump to new page with low latency
- Compatible with traditional browsers such as IE
- SEO search engine optimization
- Page open faster with server-side rendering
Order processing technology solution
- Orders are stored in both Redis and NSQ queues
- Synchronize orders to the database by synchronizing the Worker
- After the user pays the order, the order state machine drives the order status of the change database.
Order storage solution
Order data, the amount of data is too large, we use the following database sharding storage:
Website concurrency analysis
Function: Confirm Order Request
Hardware configuration: CPU(Intel Xeon Gold 6149) 2 cores、3.1GHz, memory 4G
Using the wrk stress test framework to perform a stress test of 10,000 and 20,000 concurrent amounts on a single server, you can get the average delay time and system throughput.
concurrent volume |
average delay time |
throughput |
total requests |
10000 |
0.5s |
3407 |
34407 |
20000 |
1.38s |
4795 |
48439 |