CA-Dedupe: content-aware deduplication in SSDs

  • PDF / 1,094,303 Bytes
  • 21 Pages / 439.37 x 666.142 pts Page_size
  • 110 Downloads / 201 Views

DOWNLOAD

REPORT


CA‑Dedupe: content‑aware deduplication in SSDs Ramin Gholami Taghizadeh1 · Reza Gholami Taghizadeh1 · Fahimeh Khakpash2 · Mohammadreza Binesh Marvasti1   · Seyyed Amir Asghari1

© Springer Science+Business Media, LLC, part of Springer Nature 2020

Abstract Flash memories have been around for many years because of their high performance compared to HDDs. But flash memories have a limited lifespan, and they will wear prematurely if used in write-intensive usages. Solutions such as wear leveling, compression and deduplication have been proposed to address this issue. Deduplication is a proper way to improve flash memories’ lifespan, but deduplication methods proposed in previous works usually impose a significant delay on write operations. This paper provides an intelligent method for data deduplication on flash memories which works by categorizing write requests based on their contents and types. In this scheme, calculated metadata for write requests is placed in separate categories and during deduplication procedure, the search operation is performed in one category. As a result, the proposed method improves the search delay and the deduplication rate significantly. Simulation results show that the proposed method improves delay of write operations by 32%, when compared to other methods, and achieves the deduplication rate of 69.8%. Keywords  Flash memory · SSD · Deduplication · Histogram of bytes · File type detection

* Mohammadreza Binesh Marvasti [email protected] Ramin Gholami Taghizadeh [email protected] Reza Gholami Taghizadeh [email protected] Fahimeh Khakpash [email protected] Seyyed Amir Asghari [email protected] 1

Department of Electrical and Computer Engineering, Kharazmi University, Tehran, Iran

2

Department of Industrial Engineering, Kharazmi University, Tehran, Iran



13

Vol.:(0123456789)



R. Gholami Taghizadeh et al.

1 Introduction

SDRAM Cache Cache Controller

Flash Translation Layer

RC Processor

Fig. 1  Internal components of SSD [4]

13

Flash Controller

Host Interface (SATA)

External storage of high-performance computers plays an essential role in data storage and retrieval, power consumption and system reliability. The use of silicon technologies in building non-volatile memories has led to the emergence of a new generation of storage devices called solid-state drives (SSDs). Since SSDs have been made of electronic components, they have features such as small size, low power consumption and mechanical shock resistant and are a great option for use in embedded and portable systems [1, 2]. NAND flash memory chips are one of the most common types of memory used in solid-state drives. Figure 1 shows the internal components of an SSD. In SSDs, several NAND flash memory chips use a shared channel and are connected to the flash controller module by this channel. The flash controller generates read, write and erase signals for the flash chips, and the flash memory chips connected to a channel use shared control signals. There are several data transfer chan