FMMU: A Hardware-Automated Flash Map Management Unit for Scalable Performance of NAND Flash-Based SSDs
نویسندگان
چکیده
NAND flash-based Solid State Drives (SSDs), which are widely used from embedded systems to enterprise servers, are enhancing performance by exploiting the parallelism of NAND flash memories. To cope with the performance improvement of SSDs, storage systems have rapidly adopted the host interface for SSDs from Serial-ATA, which is used for existing hard disk drives, to high-speed PCI express. Since NAND flash memory does not allow in-place updates, it requires special software called Flash Translation Layer (FTL), and SSDs are equipped with embedded processors to run FTL. Existing SSDs increase the clock frequency of embedded processors or increase the number of embedded processors in order to prevent FTL from acting as bottleneck of SSD performance, but these approaches are not scalable. This paper proposes a hardware-automated Flash Map Management Unit, called FMMU, that handles the address translation process dominating the execution time of the FTL by hardware automation. FMMU provides methods for exploiting the parallelism of flash memory by processing outstanding requests in a non-blocking manner while reducing the number of flash operations. The experimental results show that the FMMU reduces the FTL execution time in the map cache hit case and the miss case by 44% and 37%, respectively, compared with the existing software-based approach operating in 4-core. FMMU also prevents FTL from acting as a performance bottleneck for up to 32-channel, 8-way SSD using PCIe 3.0 x32 host interface.
منابع مشابه
BlueSSD: An Open Platform for Cross-layer Experiments for NAND Flash-based SSDs
In this paper we describe BlueSSD, an open platform for exploring hardware and software for NAND flash-based SSD architectures. We introduce the overall architecture of BlueSSD from a hardware and software perspective and briefly explain our design methodology. Preliminary evaluation shows that BlueSSD delivers performance comparable to commercially available SSDs.
متن کاملUsing GPUs to Forensically Recover Non-Addressable Data on Hardware-FTL-Backed NAND Storage
We present work-in-progress towards a general, parallelized method to identify leftover data in nonaddressable regions of NAND flash storage devices for which the storage interface does not allow access to all the data regions (such as with thumb drives and SSDs). Significant amounts of data can wind up in non-addressable regions through regular use of the NAND flash storage that cannot be acce...
متن کاملOne of the practices I inherited from Won
Big data real-time processing aims for faster retrieval of data and analysis. Lately, in order to accelerate real-time processing, big data platforms are trying to exploit NAND flash based storage devices, especially SSDs. NoSQL DBMSs have been used for real-time management of big data which significantly depends on index structures to efficiently manage data. Previous research about flash-awar...
متن کاملLeveraging Value Locality in Optimizing NAND Flash-based SSDs
NAND flash-based solid-state drives (SSDs) are increasingly being deployed in storage systems at different levels such as buffer-caches and even secondary storage. However, the poor reliability and performance offered by these SSDs for write-intensive workloads continues to be their key shortcoming. Several solutions based on traditionally popular notions of temporal and spatial locality help r...
متن کاملMPSearch: Multi-Path Search for Tree-based Indexes to Exploit Internal Parallelism of Flash SSDs
Big data real-time processing aims for faster retrieval of data and analysis. Lately, in order to accelerate real-time processing, big data platforms are trying to exploit NAND flash based storage devices, especially SSDs. NoSQL DBMSs have been used for real-time management of big data which significantly depends on index structures to efficiently manage data. Previous research about flash-awar...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
- CoRR
دوره abs/1704.03168 شماره
صفحات -
تاریخ انتشار 2017