Local Cache Strategy for Distributed Systems

In services with frequent read requests, such as a communication brokerage platform, introducing a cache is essential to reduce database load. In environments that scale out servers for high availability, data inconsistency between servers can occur.

Global cache (Redis) allows data sharing across servers, which is advantageous for maintaining consistency, but it incurs network traffic. On the other hand, local cache offers the advantage of fast response times without network latency and reduces dependency on external services.

In this case, Eventual Consistency is adopted for meta information that does not change frequently, and local cache is used. To achieve this, the following structure was built.

• Integrated management of local and Redis caches through CacheManager
• Type-safe cache implementation using Kotlin's sealed class
• Efficient data lookup and insertion using the getOrPut pattern

By distinguishing between local cache and Redis according to the characteristics of the data and purpose, both system performance and efficiency are secured simultaneously.