Optimizing the Cost of Garbage Collection in F2FS Using Working Set Strategy
DOI:
https://doi.org/10.54097/hset.v39i.6673Keywords:
Log Structured File System; F2FS; Garbage Collection; Hot and Cold Data Separation.Abstract
As an important part of the operating system, the file system has always received wide attention, and different types of file system architectures have been constantly proposed and practiced, this time, we mainly focus on the log structured file system (LFS) type flash friendly file system (F2FS) file system. The existing garbage collection algorithm has offered a interface for future improvement. Hence, we tried to improve existing garbage collection algorithms by implementing the working set strategy in victim select algorithm and analyze the existing deficiencies of the F2FS file system. Under the original strategy, when F2FS selects a segment, it will select the section with the fewest valid blocks. After we change it, we have completed the balance consideration of the change time, the number of valid blocks, and the proportion of changes. It is expected that the Flash device will have the longer life. By using the proposed technique, the cost of garbage collection has reduced by 10%. At the same time, based on the cold and hot data separation characteristics of the F2FS file system, we propose optimization suggestions for the cold and hot data separation strategy.
Downloads
References
Choi D, Shin D. Semantic-Aware Hot Data Selection Policy for Flash File System in Android-Based Smartphones[C]// 2013 International Conference on Parallel and Distributed Systems. IEEE, 2013: 444-445.
Lee C, Sim D, Hwang J, et al. {F2FS}: A New File System for Flash Storage[C]//13th USENIX Conference on File and Storage Technologies (FAST 15). 2015: 273-286.
Unsorted block image file system. http: //www.linux-mtd.infradead.org/ doc/ubifs.html.
Yet another flash file system. http://www. yaffs.net/.
Engel J, Mertens R. LogFS-finally a scalable flash file system[C]//12th International Linux System Technology Conference. 2005.
Kim J, Shim H, Park S Y, et al. Flashlight: A lightweight flash file system for embedded systems[J]. ACM Transactions on Embedded Computing Systems (TECS), 2012, 11(1): 1-23.
Woodhouse D. JFFS: The journalling flash file system[C]//Ottawa Linux symposium. 2001, 2001.
Gao C, Di Y, Deng A, et al. F2FS aware mapping cache design on solid state drives[C]//2018 IEEE 7th Non-Volatile Memory Systems and Applications Symposium (NVMSA). IEEE, 2018: 31-36.
Zhou Y, Wu F, Huang P, et al. An efficient page-level FTL to optimize address translation in flash memory [C]//Proceedings of the Tenth European Conference on Computer Systems. 2015: 1-16.
Min C, Kim K, Cho H, et al. SFS: random write considered harmful in solid state drives[C]//FAST. 2012, 12: 1-16.
Gupta A, Kim Y, Urgaonkar B. DFTL: a flash translation layer employing demand-based selective caching of page-level address mappings[J]. Acm Sigplan Notices, 2009, 44(3): 229-240.
Downloads
Published
Issue
Section
License

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.







