CCF-LRU: hybrid storage cache replacement strategy based on counting cuckoo filter hot-probe method

In recent years, the massive increase in the data generation has given rise to enormous challenges in storage systems. Single storage media, such as hard disk drives (HDDs) or solid state drives (SSDs), cannot meet the actual needs owing to their inherent physical characteristics. One feasible solution is to adopt a hybrid storage architecture that uses both an SSD and HDD. In this case, the management of the cache replacement strategy of the hybrid storage becomes key in improving storage performance. Based on the cuckoo filter, this study proposes a counting cuckoo filter (CCF) hot-probe method that exhibits a high space and time efficiency and supports deletion. Moreover, a CCF-least recently used (LRU) cache replacement strategy is proposed by combining CCF and the adaptive two-level LRU technique. This strategy uses CCF to identify hot data and the adaptive two-level LRU technique to manage the cache. Experimental results indicate that in comparison with traditional strategies, the cache replacement strategy combined with the hot-probe method can significantly improve cache hit ratios. Furthermore, in comparison with other cache replacement strategies that use hot-probe methods, CCF-LRU exhibits a lower time and space complexity and a higher hit ratio.