P65: CAPES: Unsupervised System Performance Tuning Using
Neural Network-Based Deep Reinforcement Learning
SessionPoster Reception
Event Type
ACM Student Research Competition
Poster
Reception
TimeTuesday, November 14th5:15pm -
7pm
LocationFour Seasons Ballroom
DescriptionParameter tuning is an important task of storage
performance optimization. Current practice usually
involves numerous tweak-benchmark cycles that are slow
and costly. To address this issue, we developed CAPES, a
model-less deep reinforcement learning-based
unsupervised parameter tuning system driven by a deep
neural network (DNN). It is designed to find optimal
values for computer systems that have tunable parameters
when human tuning can be costly and often cannot achieve
optimal performance. CAPES takes periodic measurements
of a target computer system's state, and trains a DNN
which uses Q-learning to suggest changes to the system's
current parameter values. CAPES is minimally intrusive,
and can be deployed into a production system to collect
training data and suggest tuning actions during the
system's daily operation. Evaluation of a prototype on a
Lustre file system demonstrates an increase in I/O
throughput up to 45% at saturation point.
