Reinforcement Learning-Assisted Cache Cleaning to Mitigate Long-Tail Latency in DM-SMR

DM-SMR adopts Persistent Cache (PC) to accommodate non-sequential write operations. However, the PC cleaning process induces severe long-tail latency. In this paper, we propose to mitigate the tail latency of PC cleaning by using Reinforcement Learning (RL). Specifically, a real-time lightweight Q-l...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:2021 58th ACM/IEEE Design Automation Conference (DAC) s. 103 - 108
Hlavní autoři: Pan, Yungang, Jia, Zhiping, Shen, Zhaoyan, Li, Bingzhe, Chang, Wanli, Shao, Zili
Médium: Konferenční příspěvek
Jazyk:angličtina
Vydáno: IEEE 05.12.2021
Témata:
Cleaning
Cleaning Process
Design automation
Idle Time Window
Linux
Real-time systems
Reinforcement learning
Shingled Magnetic Recording
Tail-Latency
Time factors
Timing
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:DM-SMR adopts Persistent Cache (PC) to accommodate non-sequential write operations. However, the PC cleaning process induces severe long-tail latency. In this paper, we propose to mitigate the tail latency of PC cleaning by using Reinforcement Learning (RL). Specifically, a real-time lightweight Q-learning model is built to analyze the idle window of I/O workloads, based on which PC cleaning is judiciously scheduled, thereby maximally utilizing the I/O idle window and effectively hiding the tail latency from regular requests. We implement our technique inside a Linux device driver with an emulated SMR drive. Experimental results show that our technique can reduce the tail latency by 57.65% at 99.9th percentile and the average response time by 46.11% compared to a typical SMR design.
DOI:10.1109/DAC18074.2021.9586084