# On Performance Stability in LSM-based Storage Systems (Extended Version)

## TL;DR

This paper evaluates performance stability issues in LSM-tree storage systems, focusing on write stalls and design choices to improve consistency, based on extensive experiments in Apache AsterixDB.

## Contribution

It introduces a two-phase experimental approach to assess write stalls and explores LSM merge scheduler designs to reduce stalls within I/O constraints.

## Key findings

- Identification of key factors causing write stalls.
- Effective merge scheduling strategies to minimize stalls.
- Empirical results demonstrating improved stability in AsterixDB.

## Abstract

The Log-Structured Merge-Tree (LSM-tree) has been widely adopted for use in modern NoSQL systems for its superior write performance. Despite the popularity of LSM-trees, they have been criticized for suffering from write stalls and large performance variances due to the inherent mismatch between their fast in-memory writes and slow background I/O operations. In this paper, we use a simple yet effective two-phase experimental approach to evaluate write stalls for various LSM-tree designs. We further explore the design choices of LSM merge schedulers to minimize write stalls given an I/O bandwidth budget. We have conducted extensive experiments in the context of the Apache AsterixDB system and we present the results here.