2018/11/05 by Donghyun Gouk, Gouk, Donghyun, Miryeong Kwon +13 · 8 citations
Computer Science · #Advanced Data Storage Technologies #Caching and Content Delivery #Distributed and Parallel Computing Systems #FOS: Computer and information sciences #Hardware Architecture (cs.AR) #Parallel Computing and Optimization Techniques
paper · pdf · doi:10.48550/arxiv.1811.01544
openalex publication_date 2018/11/05 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
SSDs become a major storage component in modern memory hierarchies, and SSD\nresearch demands exploring future simulation-based studies by integrating SSD\nsubsystems into a full-system environment. However, several challenges exist to\nmodel SSDs under a full-system simulations; SSDs are composed upon their own\ncomplete system and architecture, which employ all necessary hardware, such as\nCPUs, DRAM and interconnect network. Employing the hardware components, SSDs\nalso require to have multiple device controllers, internal caches and software\nmodules that respect a wide spectrum of storage interfaces and protocols. These\nSSD hardware and software are all necessary to incarnate storage subsystems\nunder full-system environment, which can operate in parallel with the host\nsystem. In this work, we introduce a new SSD simulation framework, SimpleSSD\n2.0, namely Amber, that models embedded CPU cores, DRAMs, and various flash\ntechnologies (within an SSD), and operate under the full system simulation\nenvironment by enabling a data transfer emulation. Amber also includes full\nfirmware stack, including DRAM cache logic, flash firmware, such as FTL and\nHIL, and obey diverse standard protocols by revising the host DMA engines and\nsystem buses of a popular full system simulator's all functional and timing CPU\nmodels (gem5). The proposed simulator can capture the details of dynamic\nperformance and power of embedded cores, DRAMs, firmware and flash under the\nexecutions of various OS systems and hardware platforms. Using Amber, we\ncharacterize several system-level challenges by simulating different types of\nfullsystems, such as mobile devices and general-purpose computers, and offer\ncomprehensive analyses by comparing passive storage and active storage\narchitectures.\n