A Recommendation-Based Auxiliary Caching for Mapping Record

Due to the increase in the volume of data and intensified concurrent requests, distributed caching is commonly used to manage high-concurrency requests and alleviate pressure on databases. However, there is limited research on distributed record mapping caching, and traditional caching algorithms have suboptimal resolution performance for mapping records that typically follow a long-tail distribution. To address the aforementioned issue, in this paper, we propose a recommendation-based adaptive auxiliary caching method, AC-REC, which delivers the primary cache record along with a list of additional cache records. The method uses request correlations as a basis for recommendations, customizes the number of additional cache entries provided, and dynamically adjusts the time-to-live. We conducted evaluations to compare the performance of our method against various benchmark strategies. The results show that our proposed method, as compared to the conventional LCE method, increased the cache hit ratio by an average of 20%, Moreover, this improvement is achieved while effectively utilizing the cache space. We believe that our strategy will contribute an effective solution to the related studies in both traditional network architecture and caching in paradigms like ICN.