On-line data compression in a log-structured file system
ASPLOS V Proceedings of the fifth international conference on Architectural support for programming languages and operating systems
The log-structured merge-tree (LSM-tree)
Acta Informatica
Transaction Support in a Log-Structured File System
Proceedings of the Ninth International Conference on Data Engineering
On efficient wear leveling for large-scale flash-memory storage systems
Proceedings of the 2007 ACM symposium on Applied computing
Design of flash-based DBMS: an in-page logging approach
Proceedings of the 2007 ACM SIGMOD international conference on Management of data
Design tradeoffs for SSD performance
ATC'08 USENIX 2008 Annual Technical Conference on Annual Technical Conference
Characterizing flash memory: anomalies, observations, and applications
Proceedings of the 42nd Annual IEEE/ACM International Symposium on Microarchitecture
Tracking back references in a write-anywhere file system
FAST'10 Proceedings of the 8th USENIX conference on File and storage technologies
DFS: a file system for virtualized flash storage
FAST'10 Proceedings of the 8th USENIX conference on File and storage technologies
Write endurance in flash drives: measurements and analysis
FAST'10 Proceedings of the 8th USENIX conference on File and storage technologies
OSDI'08 Proceedings of the 8th USENIX conference on Operating systems design and implementation
FAST'11 Proceedings of the 9th USENIX conference on File and stroage technologies
LazyFTL: a page-level flash translation layer optimized for NAND flash memory
Proceedings of the 2011 ACM SIGMOD International Conference on Management of data
I-CASH: Intelligently Coupled Array of SSD and HDD
HPCA '11 Proceedings of the 2011 IEEE 17th International Symposium on High Performance Computer Architecture
Beyond block I/O: Rethinking traditional storage primitives
HPCA '11 Proceedings of the 2011 IEEE 17th International Symposium on High Performance Computer Architecture
A file is not a file: understanding the I/O behavior of Apple desktop applications
SOSP '11 Proceedings of the Twenty-Third ACM Symposium on Operating Systems Principles
Object-based SCM: An efficient interface for Storage Class Memories
MSST '11 Proceedings of the 2011 IEEE 27th Symposium on Mass Storage Systems and Technologies
Delta-FTL: improving SSD lifetime via exploiting content locality
Proceedings of the 7th ACM european conference on Computer Systems
De-indirection for flash-based SSDs with nameless writes
FAST'12 Proceedings of the 10th USENIX conference on File and Storage Technologies
The bleak future of NAND flash memory
FAST'12 Proceedings of the 10th USENIX conference on File and Storage Technologies
FAST'12 Proceedings of the 10th USENIX conference on File and Storage Technologies
IEEE Communications Magazine
Hi-index | 0.00 |
Flash memory has gained in popularity as storage devices for both enterprise and embedded systems because of its high performance, low energy and reduced cost. The endurance problem of flash memory, however, is still a challenge and is getting worse as storage density increases with the adoption of multi-level cells (MLC). Prior work has addressed wear leveling and data reduction, but there is significantly less work on using the file system to improve flash lifetimes. Some common mechanisms in traditional file systems, such as journaling, metadata synchronization, and page-aligned update, can induce extra write operations and aggravate the wear of flash memory. This problem is called write amplification from file systems. In order to mitigate write amplification, we propose an object-based flash translation layer design (OFTL), in which mechanisms are co-designed with flash memory. By leveraging page metadata, OFTL enables lazy persistence of index metadata and eliminates journals while keeping consistency. Coarse-grained block state maintenance reduces persistent free space management overhead. With byte-unit access interfaces, OFTL is able to compact and co-locate the small updates with metadata to further reduce updates. Experiments show that an OFTL-based system, OFSS, offers a write amplification reduction of 47.4%-89.4% in SYNC mode and 19.8%-64.0% in ASYNC mode compared with ext3, ext2, and btrfs on an up-to-date page-level FTL.