每年專案
摘要
Some previous studies have been proposed to alleviate the impact of the frequently-updated parity data on an SMR-based RAID 5 storage system through the integration of extra hardware devices. They always utilize solid-state drives as a buffer to absorb the impact of random write operations on the SMR-based storage system performance. Unlike previous studies, this work proposes an SMR-friendly RAID 5 storage system, namely HSMR-RAID, that is developed on a pure SMR environment without any additional storage drives. The HSMR-RAID is based on a host-managed SMR architecture. The principle of HSMR-RAID is to exploit the benefits brought by data information from the host system side to put data blocks that would be invalidated at the same time in the same stripe. By this principle, we can reduce the amount of data movement and recalculate parity data so as to minimize garbage collection overhead. According to our experiments, the proposed scheme can speed up garbage collection performance by up to 29%.
原文 | ???core.languages.en_GB??? |
---|---|
主出版物標題 | Proceedings of the 38th ACM/SIGAPP Symposium on Applied Computing, SAC 2023 |
發行者 | Association for Computing Machinery |
頁面 | 294-296 |
頁數 | 3 |
ISBN(電子) | 9781450395175 |
DOIs | |
出版狀態 | 已出版 - 27 3月 2023 |
事件 | 38th Annual ACM Symposium on Applied Computing, SAC 2023 - Tallinn, Estonia 持續時間: 27 3月 2023 → 31 3月 2023 |
出版系列
名字 | Proceedings of the ACM Symposium on Applied Computing |
---|
???event.eventtypes.event.conference???
???event.eventtypes.event.conference??? | 38th Annual ACM Symposium on Applied Computing, SAC 2023 |
---|---|
國家/地區 | Estonia |
城市 | Tallinn |
期間 | 27/03/23 → 31/03/23 |
指紋
深入研究「HSMR-RAID: Enabling a Low Overhead RAID-5 System over a Host-managed Shingled Magnetic Recording Disk Array」主題。共同形成了獨特的指紋。專案
- 2 已完成
-
用於高效能機器學習框架之新興記憶體與儲存系統設計-於混合式記憶儲存平台上高效能資料索引系統設計以支援機器學習應用(1/3)
Chen, T.-Y. (PI)
1/08/22 → 31/07/23
研究計畫: Research
-