ENDURE: A Robust Tuning Paradigm for LSM Trees
Authors: Andy Huynh, Harshal A. Chaudhari, Evimaria Terzi, Manos Athanassoulis
(VLDB 2022). LSM-tree based storage engines are frequently tuned for performance with assumptions on the expected workloads, however, if future workloads in production deviates too far from the expected, these engines can experience heavy performance cliffs. We augment the tuning process to be robust to these perturbations in workloads.