1. Brief description
In fact, at this stage to come up with a distributed cache, some early, pre mainly use the local cache, I use the technology mainly ehcahe, this memory there is the basic server applications that run on top of you, that there is a big question is, not suitable for long-term storage, if long-term storage, when a large amount of data, will occupy a large part of your service memory, distributed caching with memcached and more are redis late, but I mainly use the redis.
redis distributed cache will have a series of questions, such as: cache coherency, the cache penetration / breakdown, avalanche cache, a hot spot data set failed. I will write a post-solutions for these issues to this problem.
2. A flowchart
3. Question
Most requests for cache Kang Zhu, subscriber growth, concurrent pressure will fall on the tomcat, the response is very slow. Here I did not understand understanding, complicated by a tomcat per second, and said the line to see a lot of support 150 concurrent default, it can be changed to 250 concurrent / sec. Individuals really want to verify the number of concurrent. So I study a little, links.
4. optimizations
-
On the Tomcat server or increase the local cache in the same JVM
-
Increase in external distributed cache
-
Caching hot data and static html page
By caching the most requested prior to read and write database interception off, can effectively improve the access speed of the application.