Magpie: Automatically Tuning Static Parameters for Distributed File Systems using Deep Reinforcement Learning.
Houkun ZhuDominik ScheinertLauritz ThamsenKain Kordian GontarskaOdej KaoPublished in: CoRR (2022)
Keyphrases
- file system
- reinforcement learning
- scalable distributed
- data transfer
- distributed data
- multi agent
- fine tuning
- distributed systems
- parameter settings
- maximum likelihood
- parameter values
- cooperative
- multi tiered
- storage systems
- flash memory
- distributed environment
- markov decision processes
- machine learning
- application specific
- access patterns
- communication cost
- search tools
- function approximation
- optimal policy
- fine tuned
- tuning parameters
- continuous media
- state space