• ETRI Journal
• /
• v.39 no.6
• /
• pp.820-831
• /
• 2017

File systems and applications try to implement their own update protocols to guarantee data consistency, which is one of the most crucial aspects of computing systems. However, we found that the storage devices are substantially under-utilized when preserving data consistency because they generate massive storage write traffic with many disk cache flush operations and force-unit-access (FUA) commands. In this paper, we present DJFS (Delta-Journaling File System) that provides both a high level of performance and data consistency for different applications. We made three technical contributions to achieve our goal. First, to remove all storage accesses with disk cache flush operations and FUA commands, DJFS uses small-sized NVRAM for a file system journal. Second, to reduce the access latency and space requirements of NVRAM, DJFS attempts to journal compress the differences in the modified blocks. Finally, to relieve explicit checkpointing overhead, DJFS aggressively reflects the checkpoint transactions to file system area in the unit of the specified region. Our evaluation on TPC-C SQLite benchmark shows that, using our novel optimization schemes, DJFS outperforms Ext4 by up to 64.2 times with only 128 MB of NVRAM.

• JSTS:Journal of Semiconductor Technology and Science
• /
• v.17 no.1
• /
• pp.1-6
• /
• 2017

This paper presents an optimized adoption of NVM for the storage system of heterogeneous applications. Our analysis shows that a bulk of I/O does not happen on a single storage partition, but it is varied significantly for different application categories. In particular, journaling I/O accounts for a dominant portion of total I/O in DB applications like OLTP, whereas swap I/O accounts for a large portion of I/O in graph visualization applications, and file I/O accounts for a large portion in web browsers and multimedia players. Based on these observations, we argue that maximizing the performance gain with NVM is not obtained by fixing it as a specific storage partition but varied widely for different applications. Specifically, for graph visualization, DB, and multimedia player applications, using NVM as a swap, a journal, and a file system partitions, respectively, performs well. Our optimized adoption of NVM improves the storage performance by 10-61%.

• The KIPS Transactions:PartA
• /
• v.14A no.1 s.105
• /
• pp.39-46
• /
• 2007

Recently flash memory is widely accepted as a storage devise of embedded systems for portability and performance reasons. Flash memory has many distinguishing features compared to legacy magnetic disks. Especially, a file system for flash memory usually assumes the form of log-structured file system and it employs garbage collector accordingly. Since the garbage collector can greatly affect the performance of file system, it should be designed carefully considering flash memory features. In this paper, we suggest a new garbage collector for existing JFFS2 (Journaling Flash File System II) file system. By extensive performance evaluation, we show that the proposed garbage collector achieves improved performance in terms of flash memory consumption rate, increased flash memory life time, and improved wear-leveling.

• Journal of Korea Society of Digital Industry and Information Management
• /
• v.6 no.2
• /
• pp.107-118
• /
• 2010

This paper proposes a digital forensic method to a file creation transaction using a journal file($LogFile) on NTFS File System. The journal file contains lots of information which can help recovering the file system when system failure happens, so knowledge of the structure is very helpful for a forensic analysis. The structure of the journal file, however, is not officially opened. We find out the journal file structure with analyzing the structure of log records by using reverse engineering. We show the digital forensic procedure extracting information from the log records of a sample file created on a NTFS volume. The related log records are as follows: bitmap and segment allocation information of MFT entry, index entry allocation information, resident value update information($FILE_NAME, $STANDARD_INFORMATION, and INDEX_ALLOCATION attribute etc.).

• The KIPS Transactions:PartA
• /
• v.8A no.4
• /
• pp.375-384
• /
• 2001

The shared disk file systems use a technique known as file system journaling to support recovery of metadata on the SAN(Storage Area Network). In the existing journaling technique, the metadata that is dirtied by one host must be updated to disk space before some hosts access it. The system performance is decreased because the disk access number is increased. In this paper, we describe a new recovery technique using a global buffer to decrease disk I/O. It transmits the dirtied metadata into the other hosts through Fibre Channel network on the SAN instead of disk I/O and supports recovery of a critical data by journaling a data as well as metadata.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.16 no.6
• /
• pp.177-181
• /
• 2016

This paper analyzes the host cache effectiveness in full virtualization, particularly associated with journaling of guests. We observe that the journal access of guests degrades cache performance significantly due to the write-once access pattern and the frequent sync operations. To remedy this problem, we design and implement a novel caching policy, called PDC (Pollution Defensive Caching), that detects the journal accesses and prevents them from entering the host cache. The proposed PDC is implemented in QEMU-KVM 2.1 on Linux 4.14 and provides 3-32% performance improvement for various file and I/O benchmarks.

• Journal of KIISE
• /
• v.44 no.4
• /
• pp.363-373
• /
• 2017

Ext4 file system is widely used in various computing environments such as those of the PC, the server, and the Linux-based embedded system. Ext4, which uses a buffer for block I/O, provides fsync() system call to applications to guarantee the consistency of a specific file. A log of the analytical studies regarding the operation of Ext4 and the improvement of its performance has been compiled, but it has not been studied in detail in terms of kernel versions. We figure out that the behavior of fsync() system call is different depending on the kernel version. Between the kernel versions of 3.4.0 and 4.7.2, 3.4.0, 3.8.0, and 4.6.2 showed behavioral differences regarding the fsync() system call. The latency of fsync() in kernel 3.4.0 is longer than that of the more-advanced 3.7.10; meanwhile, the characteristics of 3.8.0 enabled the disruption of the Ext4 journaling order, but the ordered defect was solved with 4.6.2.

• Journal of IKEEE
• /
• v.23 no.1
• /
• pp.80-90
• /
• 2019

Even though Non-Volatile Memory (NVM) is used for data storage, a page table should be built to access data in it. And this observation leads us to the Persistent Page Table (PPT) scheme that keeps the page table in NVM persistently. By the way, processors have different page table structures and really operational page table cannot be built without virtual and physical addresses of NVM. However, those addresses are determined dynamically when NVM storage is attached to the system. Thus, the PPT should have system-independent and also address-independent structure and really working system-dependent page table should be built from the PPT. Moreover, entries of PPT should be updated atomically and, in this paper, we describe the design of PPT that meets those requirements. And we investigate how file systems can decrease the journaling overhead with the swap operation, which is a new operation created by the PPT. We modified the Ext4 file system in Linux and experiments conducted with Filebench workloads show that the swap operation enhances file system performance up to 60%.

• The KIPS Transactions:PartA
• /
• v.8A no.4
• /
• pp.367-374
• /
• 2001

Recently, LINUX cluster file systems based on the storage area network (SAN) have been developed. In those systems, without using a central file server, multiple clients sharing the whole disk storage through Fibre Channel can freely access disk storage and act as file servers. Accordingly, they can offer advantages such as availability, load balancing, and scalability. In this paper, we describe metadata management schemes designed for a new SAN-based LINUX cluster file system. First, we present a new inode structure which is better than previous ones in disk block access time. Second, a new directory structure which uses extendible hashing is described. Third, we describe a novel scheme to manage free disk blocks, which is suitable for very large file systems. Finally, we present how we handle metadata journaling. Through performance evaluation, we show that our proposed schemes have better performance than previous ones.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.49 no.4
• /
• pp.1-8
• /
• 2012

The NTFS journaling file($LogFile) is used to keep the file system clean in the event of a system crash or power failure. The operation on files leaves large amounts of information in the $LogFile. Despite the importance of a journal file as a forensic evidence repository, its structure is not well documented. The researchers used reverse engineering in order to gain a better understanding of the log record structures of address parts, and utilized the address for identifying object files to gain forensic information.