Efficient Compactions Between Storage Tiers with PrismDB

TL;DR

PrismDB is a novel key-value store that efficiently manages data across fast and low-cost storage tiers, achieving significant performance improvements and cost efficiency through a new multi-tiered compaction algorithm.

Contribution

The paper introduces PrismDB, a storage engine that effectively utilizes two different NVMe storage technologies and proposes a new algorithm for multi-tiered data compaction.

Findings

3.3x higher average throughput compared to RocksDB on tiered storage.

2x better read tail latency than RocksDB.

Achieves Pareto-efficient balance between performance and cost.

Abstract

In recent years, emerging storage hardware technologies have focused on divergent goals: better performance or lower cost-per-bit. Correspondingly, data systems that employ these technologies are typically optimized either to be fast (but expensive) or cheap (but slow). We take a different approach: by architecting a storage engine to natively utilize two tiers of fast and low-cost storage technologies, we can achieve a Pareto-efficient balance between performance and cost-per-bit. This paper presents the design and implementation of PrismDB, a novel key-value store that exploits two extreme ends of the spectrum of modern NVMe storage technologies (3D XPoint and QLC NAND) simultaneously. Our key contribution is how to efficiently migrate and compact data between two different storage tiers. Inspired by the classic cost-benefit analysis of log cleaning, we develop a new algorithm for multi-tiered storage compaction that balances the benefit of reclaiming space for hot objects in fast storage with the cost of compaction I/O in slow storage. Compared to the standard use of RocksDB on flash in datacenters today, PrismDB's average throughput on tiered storage is 3.3$\times$ faster and its read tail latency is 2$\times$ better, using equivalently priced hardware.