Redis cache update strategy, cache penetration, avalanche, breakdown problems

Preface

This article is learned by me from dark horse programmer B Stop video tutorial （https://www.bilibili.com/video/BV1cr4y1671t） Notes made in the process of , Mainly records Redis Some ideas about caching , No code involved .

Redis cache

effect

• Reduce back-end load
• Improve reading and writing efficiency , Reduce response time

cost

• Data consistency costs
• Code maintenance costs
• O & M costs

Cache update strategy

1. Memory obsolescence

   • Automatically eliminate when the content is insufficient

   • Default on

   • You don't have to maintain it yourself

   • Poor consistency

   • Low maintenance cost

2. Time out culling

   • add to TTL Automatically delete when due

   • General consistency

   • Low maintenance cost

3. Active update

   • Write business logic , When modifying the database , Update cache

   • Good consistency

   • Maintenance costs are high

Select according to business scenario ：

Low consistency requirements ： Use Redis Built in memory elimination mechanism .

High consistency requirements ： Active update , And enable timeout culling ,

1. Read operations

   • Cache hits return directly

   • If not, query the database , And write to the cache

2. Write operations

   • Write the database first , Then delete the cache

   • To ensure the atomicity of database and cache operations ( Monomer applications can use @Transactional Open transaction )
     

     

Cache penetration

Cache penetration means that the data requested by the client does not exist in the cache or in the database , This cache will never take effect , These requests will all go to the database .

There are two common solutions ：

1. Caching empty objects
   • advantage ： Implement a simple , Convenient maintenance
   • shortcoming ： Extra memory consumption 、 May cause short-term inconsistencies
2. Bloon filtration （ be based on bitmap Realization ）
   • advantage ： Less memory , There is no excess key
   • shortcoming ： The implementation is complex 、 There is a possibility of misjudgment

（ On the left is the schematic diagram of caching empty objects , On the right is the schematic diagram of bulon filtration ）

Cache avalanche

Cache avalanche refers to a large number of caches at the same time key Simultaneous failure or Redis The service outage , Cause a large number of requests to reach the database , It brings a lot of pressure .

（ On the left is a large number of cache invalidation diagram , The right side is redis Downtime diagram ）

Solution ：

• To different Key Of TTL add to Random value
• utilize Redis colony Improve service availability
• Add Degraded current limiting Strategy
• Add to business Multi level cache （ Browser cache 、Nginx cache 、Redis cache 、JVM Local cache 、 Databases, etc ）

Cache breakdown

Cache breakdown is also called hot spot Key problem , It is a highly concurrent access and complex cache reconstruction business key All of a sudden it failed , Countless requests for access will have a huge impact on the database in an instant .

There are two common solutions ：

• Logical expiration

  Not set up TTL, Instead, add a expire Field , When the field time reaches , Update data .

  • advantage
    • Threads do not need to wait , Good performance
  • shortcoming
    • There is no guarantee of consistency
    • There is additional memory consumption
    • The implementation is complex

• The mutex

  • advantage
    • No additional memory consumption
    • Guarantee consistency
    • Implement a simple
  • shortcoming
    • Thread needs to wait , Performance is affected
    • There may be a deadlock risk
      （ On the left is the schematic diagram of mutual exclusion , Schematic diagram of logical expiration of the right side wall ）
      

Solve the cache breakdown problem based on mutual exclusion

Solve the cache breakdown problem based on logical expiration

Postscript

u1s1, This tutorial is really good , And it's 2022 New version of video , It is highly recommended that you have a look .