{"title":"An FTL-Aware Host System Alleviating Severe Long Latency of NAND Flash-based Storage","authors":"Jung-Hoon Kim","doi":"10.1109/RTCSA52859.2021.00029","DOIUrl":null,"url":null,"abstract":"A flash translation layer (FTL) of NAND flash-based storage occasionally conducts a garbage collection causing long-wait situations to a host system. To avert that long latency, the host system using a black-box model has chased the storage performance. On the other hand, other storage systems have required expensive storage hardware or depended on idle timing. This paper proposes an FTL-aware host system extracting the FTL metadata from the storage during runtime to overcome these constraints. Furthermore, this novel host system predicts the garbage collection and directs the storage to lower the long latency in advance. The experiment results show the alleviation of the long latency against the legacy host system.","PeriodicalId":38446,"journal":{"name":"International Journal of Embedded and Real-Time Communication Systems (IJERTCS)","volume":"16 1","pages":"189-194"},"PeriodicalIF":0.5000,"publicationDate":"2021-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"2","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"International Journal of Embedded and Real-Time Communication Systems (IJERTCS)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/RTCSA52859.2021.00029","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q4","JCRName":"COMPUTER SCIENCE, SOFTWARE ENGINEERING","Score":null,"Total":0}
引用次数: 2
Abstract
A flash translation layer (FTL) of NAND flash-based storage occasionally conducts a garbage collection causing long-wait situations to a host system. To avert that long latency, the host system using a black-box model has chased the storage performance. On the other hand, other storage systems have required expensive storage hardware or depended on idle timing. This paper proposes an FTL-aware host system extracting the FTL metadata from the storage during runtime to overcome these constraints. Furthermore, this novel host system predicts the garbage collection and directs the storage to lower the long latency in advance. The experiment results show the alleviation of the long latency against the legacy host system.