[Spring] Redis as a cache memory

Redis Study Notes

---

Chapter 1 - Redis & RedisTemplate

• Redis is a fast, memory-based database.
• To use Redis in a Java-based Spring Boot project, something is needed to bridge the connection between Java and Redis.
  • RedisTemplate is what fulfills this role.
  • RedisTemplate can be thought of as a translator between Java and Redis.
  • It enables communication by mutually converting Java commands into Redis commands, and vice versa.

---

Chapter 2 - Redis Data Storage & Serialization

Storage Format

• Redis stores all data in key-value format.
• All data types are stored as strings.

Serialization (Java → Redis)

• Storing a Java object in Redis requires parsing it into string format.
• This process is called serialization, and it parses the object into JSON format.
• Different serialization tools are applied to keys and values.
  • Key : Uses StringRedisSerializer
  • Value : Varies depending on whether the type is string or hash.

Deserialization (Redis → Java)

• Restoring JSON string data from Redis back into a Java object requires a deserialization tool.
• By registering Jackson2JsonRedisSerializer or GenericJackson2JsonRedisSerializer as the value serializer during configuration, RedisTemplate will automatically deserialize and return the data as a Java object.
• Type casting is performed after that.

---

Chapter 3 - Key Naming Conventions

• Since all data in Redis is key-based, it is vulnerable to overwriting.
• To prevent unintended value overwrites caused by duplicate key names, naming conventions exist for identification purposes.
  • Since these are conventions, the specific rules may vary by organization.

Basic Format : {domain}:{resource}:{identifier}

• The colon (:) is a common agreement in Redis for expressing hierarchical structure.
• Keys are typically managed as static final constants.

---

Chapter 4 - Cache Characteristics & the Need for TTL

Characteristics of Cache Memory

• Cache memory refers to memory with small capacity and volatility, but extremely fast data access.
• The reason cache is fast is because data is stored in RAM. Small capacity and volatility are resulting characteristics of RAM, not the cause of its speed.
• Redis is primarily used to serve as this kind of cache memory.

Why Cache Needs to Be Cleared Periodically

• To guarantee fast cache lookup performance, it is necessary to periodically clear the cache so that too much data does not accumulate.
• TTL (Time To Live) is a Redis feature that sets an expiration time on each key, and is the primary mechanism for automating this cache-clearing process.

The Need for Cache Synchronization

• Cache serves as a supplement to the database.
• Since cached data is premised on being a copy of the database's values, any discrepancy between a change in the database and the value in cache must be reflected immediately — this is referred to as cache synchronization.

Trade-offs in Cache Synchronization Frequency

• If cache synchronization occurs too frequently, cache performance degrades.
  • If the cache's stored values are not trusted and the database is queried regardless, there is no point in using cache at all.
• On the other hand, the appropriate frequency of cache synchronization varies depending on the nature of the data and users' information retrieval patterns.
• Since there is no single correct answer, it is necessary to tune the cache synchronization frequency according to the needs of the business and the characteristics of the data.