A Hierarchical Storage Mechanism for Hot and Cold Data Based on Temperature Model

The storage of hot and cold data play a crucial role in improving data access efficiency and reducing storage expenses. This paper proposed a temperature model to quantify the real-time hotness of data. Based on the temperature model, we proposed a hierarchical storage mechanism for hot and cold data, managing dynamic data migration among local cold database, local hot database, and remote cold database. Experimental results show the advantages of the proposed method in terms of hot data hit rate, hot data hit rate for key data, migration count, and average response time. It can improve data access performance and the satisfaction of important users, and significantly reduce expenses.