Notes on cache sharing meeting

Time:2020-12-14

cache

principle

classification

CDN cache

QBs (number of visits per second) caches static resources as much as possible

Reverse proxy caching

Distributed cache

Hot data

Local application cache

Message middleware

Problems encountered with caching

cache

uniformity

avalanche

of large numberCache data expiration, invalidation (expired, old data)

Solution: if the first request finds that the cache fails, it will directly request the database and extend the cache. The same request continues to access the cache

High availability

pierce through

The cache data does not exist
There is no K, a large number of queries do not exist in the cache and database, and the flag value = null
If there are a large number of nonexistent K, set a large number of value = null to waste memory.
LRU cache alarm
K regular XX filter filtering garbage request rookie

In depth understanding of cookie and session mechanism

The difference between session and cookie

Mark users

Session server

Too much session data

Cookie client

The data is too long

security

Cross domain cookie synchronization

Third party cookie requested by other browsers Ajax (browser disabled)

One party cookie user visits a.com to detect the login, and the login information is stored in the background jump (Taobao tmall) jump.a.com sub domain 302
Too many domains?

One party / three parties are decided by the client

diamond

Manage configuration

Disaster recovery