Jiarong Liu , Tianyu Wang , Xiaowei Chen , Chao Li , Zhaoyan Shen , Zhiyong Zhang
{"title":"H2-RAID:利用统一的 SSD 和 HDD 混合架构提高 SSD RAID 的可靠性","authors":"Jiarong Liu , Tianyu Wang , Xiaowei Chen , Chao Li , Zhaoyan Shen , Zhiyong Zhang","doi":"10.1016/j.micpro.2023.104993","DOIUrl":null,"url":null,"abstract":"<div><p><span>With the increasing development of SSD (Solid-State Drives) technology, SSD RAID (Redundant Arrays of Independent Disks) has been widely deployed in enterprise data centers. However, the inherent write endurance issue of SSD seriously affects the reliability of the array. Meanwhile, compared with conventional HDD-based RAID, SSD RAID exhibits very different failure characteristics, such as correlated failure (Balakrishnan et al., 2010) under RAID-5. In this paper, we present a Hybrid High reliability RAID architecture, named H</span><span><math><msup><mrow></mrow><mrow><mn>2</mn></mrow></msup></math></span>-RAID, by equipping each SSD with an extra HDD as the backup to improve the reliability of SSD RAID. Considering the relatively longer write latency of HDD, in H<span><math><msup><mrow></mrow><mrow><mn>2</mn></mrow></msup></math></span><span><span>-RAID, we first propose an HDD-aware backup mechanism to smartly aggregate random writes into sequential writes to decrease performance degradation. In addition, to cope with the scenarios of SSD failure, an HDD-aware reconstruction method is further proposed to guarantee the reliability and the online </span>transaction processing performance. We build a novel Markov process-based mathematical model to analyze the reliability of different architectures, and the theoretical results prove the reliability of H</span><span><math><msup><mrow></mrow><mrow><mn>2</mn></mrow></msup></math></span>-RAID is much higher than that of traditional SSD RAID. To more accurately evaluate the performance influence of HDD on H<span><math><msup><mrow></mrow><mrow><mn>2</mn></mrow></msup></math></span>-RAID, we develop a simulator based on Disksim and the experimental results show H<span><math><msup><mrow></mrow><mrow><mn>2</mn></mrow></msup></math></span>-RAID significantly increases the reliability compared with SSD array (under RAID-5) while with little performance loss on average.</p></div>","PeriodicalId":49815,"journal":{"name":"Microprocessors and Microsystems","volume":"105 ","pages":"Article 104993"},"PeriodicalIF":1.9000,"publicationDate":"2023-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"H2-RAID: Improving the reliability of SSD RAID with unified SSD and HDD hybrid architecture\",\"authors\":\"Jiarong Liu , Tianyu Wang , Xiaowei Chen , Chao Li , Zhaoyan Shen , Zhiyong Zhang\",\"doi\":\"10.1016/j.micpro.2023.104993\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><p><span>With the increasing development of SSD (Solid-State Drives) technology, SSD RAID (Redundant Arrays of Independent Disks) has been widely deployed in enterprise data centers. However, the inherent write endurance issue of SSD seriously affects the reliability of the array. Meanwhile, compared with conventional HDD-based RAID, SSD RAID exhibits very different failure characteristics, such as correlated failure (Balakrishnan et al., 2010) under RAID-5. In this paper, we present a Hybrid High reliability RAID architecture, named H</span><span><math><msup><mrow></mrow><mrow><mn>2</mn></mrow></msup></math></span>-RAID, by equipping each SSD with an extra HDD as the backup to improve the reliability of SSD RAID. Considering the relatively longer write latency of HDD, in H<span><math><msup><mrow></mrow><mrow><mn>2</mn></mrow></msup></math></span><span><span>-RAID, we first propose an HDD-aware backup mechanism to smartly aggregate random writes into sequential writes to decrease performance degradation. In addition, to cope with the scenarios of SSD failure, an HDD-aware reconstruction method is further proposed to guarantee the reliability and the online </span>transaction processing performance. We build a novel Markov process-based mathematical model to analyze the reliability of different architectures, and the theoretical results prove the reliability of H</span><span><math><msup><mrow></mrow><mrow><mn>2</mn></mrow></msup></math></span>-RAID is much higher than that of traditional SSD RAID. To more accurately evaluate the performance influence of HDD on H<span><math><msup><mrow></mrow><mrow><mn>2</mn></mrow></msup></math></span>-RAID, we develop a simulator based on Disksim and the experimental results show H<span><math><msup><mrow></mrow><mrow><mn>2</mn></mrow></msup></math></span>-RAID significantly increases the reliability compared with SSD array (under RAID-5) while with little performance loss on average.</p></div>\",\"PeriodicalId\":49815,\"journal\":{\"name\":\"Microprocessors and Microsystems\",\"volume\":\"105 \",\"pages\":\"Article 104993\"},\"PeriodicalIF\":1.9000,\"publicationDate\":\"2023-12-30\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Microprocessors and Microsystems\",\"FirstCategoryId\":\"94\",\"ListUrlMain\":\"https://www.sciencedirect.com/science/article/pii/S0141933123002387\",\"RegionNum\":4,\"RegionCategory\":\"计算机科学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q3\",\"JCRName\":\"COMPUTER SCIENCE, HARDWARE & ARCHITECTURE\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Microprocessors and Microsystems","FirstCategoryId":"94","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S0141933123002387","RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"COMPUTER SCIENCE, HARDWARE & ARCHITECTURE","Score":null,"Total":0}
H2-RAID: Improving the reliability of SSD RAID with unified SSD and HDD hybrid architecture
With the increasing development of SSD (Solid-State Drives) technology, SSD RAID (Redundant Arrays of Independent Disks) has been widely deployed in enterprise data centers. However, the inherent write endurance issue of SSD seriously affects the reliability of the array. Meanwhile, compared with conventional HDD-based RAID, SSD RAID exhibits very different failure characteristics, such as correlated failure (Balakrishnan et al., 2010) under RAID-5. In this paper, we present a Hybrid High reliability RAID architecture, named H-RAID, by equipping each SSD with an extra HDD as the backup to improve the reliability of SSD RAID. Considering the relatively longer write latency of HDD, in H-RAID, we first propose an HDD-aware backup mechanism to smartly aggregate random writes into sequential writes to decrease performance degradation. In addition, to cope with the scenarios of SSD failure, an HDD-aware reconstruction method is further proposed to guarantee the reliability and the online transaction processing performance. We build a novel Markov process-based mathematical model to analyze the reliability of different architectures, and the theoretical results prove the reliability of H-RAID is much higher than that of traditional SSD RAID. To more accurately evaluate the performance influence of HDD on H-RAID, we develop a simulator based on Disksim and the experimental results show H-RAID significantly increases the reliability compared with SSD array (under RAID-5) while with little performance loss on average.
期刊介绍:
Microprocessors and Microsystems: Embedded Hardware Design (MICPRO) is a journal covering all design and architectural aspects related to embedded systems hardware. This includes different embedded system hardware platforms ranging from custom hardware via reconfigurable systems and application specific processors to general purpose embedded processors. Special emphasis is put on novel complex embedded architectures, such as systems on chip (SoC), systems on a programmable/reconfigurable chip (SoPC) and multi-processor systems on a chip (MPSoC), as well as, their memory and communication methods and structures, such as network-on-chip (NoC).
Design automation of such systems including methodologies, techniques, flows and tools for their design, as well as, novel designs of hardware components fall within the scope of this journal. Novel cyber-physical applications that use embedded systems are also central in this journal. While software is not in the main focus of this journal, methods of hardware/software co-design, as well as, application restructuring and mapping to embedded hardware platforms, that consider interplay between software and hardware components with emphasis on hardware, are also in the journal scope.