Enhancing Redis Cache Efficiency Based on Dynamic TTL and Adaptive Eviction Mechanism

Common caching strategies used in Redis are often inefficient when dealing with complex data structures, leading to inefficient memory usage and degraded system performance. This work presents the idea of decomposing complex data structures into simpler sub-elements and a custom caching strategy is applied to each of them. This custom caching strategy is based on dynamic Time-To-Live (TTL) determination and an assessment of real-time access patterns to more accurately balance data retention and eviction. It determines the need for eviction based on the combination of the weight of each key along with the advantages of the Least Recently Used (LRU) and Least Frequently Used (LFU) algorithms. Developed using C# on the.NET platform with the StackExchange. Redis library, the suggested solution aims to improve cache hit rates and reduce memory overhead in high-load scenarios. The results of the experiment confirm the promising nature of the developed adaptive strategy.