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

Abstract

After the introduction of Persistent Memory in the form of Intel's Optane DC Persistent Memory on the market in 2019, it has found its way into manifold applications and systems. As Google and other cloud infrastructure providers are starting to incorporate Persistent Memory into their portfolio, it is only logical that cloud applications have to exploit its inherent properties. Persistent Memory can serve as a DRAM substitute, but guarantees persistence at the cost of compromised read/write performance compared to standard DRAM. These properties particularly affect the performance of index structures, since they are subject to frequent updates and queries. However, adapting each and every index structure to exploit the properties of Persistent Memory is tedious. Hence, we require a general technique that hides this access gap, e.g., by using DRAM caching strategies. To exploit Persistent Memory properties for analytical index structures, we propose selective caching. It is based 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 evaluate selective caching on the OLAP-optimized main-memory index structure Elf, because its memory layout allows for an easy caching. Our experiments show that if configured well, selective caching with a suitable replacement strategy can keep pace with pure DRAM storage of Elf while guaranteeing persistence. These results are also reflected when selective caching is used for parallel workloads.

Selective caching: a persistent memory approach for multi-dimensional index structures

“…The competition of PMem and DRAM for free space in the CPU caches is still included here. In our initial results [17], we have shown that dynamic caching definitely gives benefits when executing the same set of queries twice (selective warm-up). Now, we query different points or ranges, but still keep a certain skewness and find that dynamic caching is no longer that significantly better.…”

“…Next, we compare the throughput of the three query types for the approaches over a time course expressed in queries run up to that point. In contrast to our prior work [17], the system and caches are not warmed up before to show how the setup and its performance evolves over time. We show the results for the persistent Elf without DRAM caching, the corresponding best dynamic caching, static caching of the first x dimension levels as well as the combination of dynamic and static caches.…”

“…In this paper, we extend our proposed approach, selective caching [17], which was a first step to exploit PMem for multi-dimensional index structures. As a representative, we use Elf [4], a multi-dimensional index structure.…”

“…• As an extension to our prior work [17], we implement several replacement strategies for dynamic caching (Sect. 4.2).…”

Designing an Event Store for a Modern Three-layer Storage Hierarchy

“…With the LB + -Tree [22] the authors refine the concept for Intel's DCPMMs by utilizing multi-256-byte nodes and limiting the number of PMem line writes. In [13] it is shown that the selective persistence concept of the FPTree is similarly applicable to other data structures. The use case here is a persistent trie-like structure enhanced with various DRAM caching strategies.…”