Selective Caching: A Persistent Memory Approach for Multi-Dimensional Index Structures

Since the proposal of Persistent Memory, research has focused on tuning a variety of data management problems to the inherent properties of Persistent Memory-namely persistence but also compromised read/write performance. These properties particularly affect the performance of index structures, since they are subject to frequent updates and queries. Nevertheless, the main research focuses on adapting B-Trees and its derivatives to Persistent Memory properties, aiming to reach DRAM processing speed exploiting the persistence property of Persistent Memory. However, most of the found techniques for B-Trees are not directly applicable to other tree-based index structures or even multi-dimensional index structures. To exploit Persistent Memory properties for arbitrary index structures, we propose selective caching. It bases on a mixture of dynamic and static caching of tree nodes in DRAM to reach near-DRAM access speeds for index structures. In this paper, we investigate the opportunities as well as limitations of selective caching on the OLAP-optimized main-memory index structure Elf. Our experiments show that selective caching is keeping up with pure DRAM storage of Elf while guaranteeing persistence.