What is the core design of Dadi's performance advantage in the cache hit scenario?

DADI Performance advantages in cache hit scenario , What is the core design ？

Take the answer 1：

1、 Short circuit reading , Read shared memory directly , Avoid passage IPC read ; 2、 Whether the cache hits the data structure , It's also in shared memory . adopt reference count, combination robust mutex To ensure the safety of multi thread data sharing ; 3、 Disk read ,100us,+ 27us About equal to the disk read itself rt,IPC go shm signal communication , No local socket signal communication . 4、 Very low resource use . Memory ：DADI Service The memory used is 100~200M, The reason is based on shared memory IPC Realization ,hash Data structures such as tables , Avoid memory expansion under multi process architecture , Streamlined coding ,1 Memory pages 16k Corresponding 4byte Management structure ; CPU：Local DADI Service When the disk is full, the single core CPU Use 20% about .CPU Used in SDK here ,SDK And Local DADI Service There is little communication .

版权声明
本文为[Alibaba cloud Q & A]所创，转载请带上原文链接，感谢
https://yzsam.com/2022/04/202204211136312565.html

当前位置：网站首页>What is the core design of Dadi's performance advantage in the cache hit scenario?

What is the core design of Dadi's performance advantage in the cache hit scenario?

边栏推荐

猜你喜欢

随机推荐