LSM Trees in Adversarial Environments

TL;DR

This paper investigates how adversarial workloads can significantly degrade read performance in LSM trees by affecting Bloom Filter accuracy, and proposes methods to improve resilience in popular key-value stores.

Contribution

It introduces adversarial models and security definitions for LSM trees, and implements resilience strategies in LevelDB and RocksDB to mitigate performance degradation.

Findings

Up to 800% increase in read latency under adversarial workloads

Adversarial models and security definitions for LSM stores

Resilience techniques reduce performance degradation

Abstract

The Log Structured Merge (LSM) Tree is a popular choice for key-value stores that focus on optimized write throughput while maintaining performant, production-ready read latencies. To optimize read performance, LSM stores rely on a probabilistic data structure called the Bloom Filter (BF). In this paper, we focus on adversarial workloads that lead to a sharp degradation in read performance by impacting the accuracy of BFs used within the LSM store. Our evaluation shows up to $800\%$ increase in the read latency of lookups for popular LSM stores. We define adversarial models and security definitions for LSM stores. We implement adversary resilience into two popular LSM stores, LevelDB and RocksDB. We use our implementations to demonstrate how performance degradation under adversarial workloads can be mitigated.