PrismDB: Read-aware Log-structured Merge Trees for Heterogeneous Storage

In recent years, emerging hardware storage technologies have focused on divergent goals: better performance or lower cost-per-bit of storage. Correspondingly, data systems that employ these new technologies are optimized either to be fast (but expensive) or cheap (but slow). We take a different approach: by combining a heterogeneous set of fast and low-cost storage technologies within the same system, we can achieve a Pareto-efficient balance between performance and cost-per-bit. This paper presents the design and implementation of PrismDB, a novel log-structured merge tree key-value store that exploits the full spectrum of heterogeneous storage technologies simultaneously. We introduce the notion of "read-awareness" to log-structured merge trees, which allows hot objects to be pinned to faster storage, achieving better tiering and hot-cold separation of objects. Compared to Mutant, a prior key-value store for heterogeneous storage, and RocksDB, PrismDB can achieve up to 5.8$\times$ and 5.1$\times$ higher throughput (respectively), reduce read tail latency by 10$\times$ and 10.7$\times$, and reduce update latency by 10.3$\times$ and 9$\times$.