ABSTRACT
File system I/O is increasingly becoming a performance bottleneck in large distributed computer systems. This is due to the increased file I/O demands of new applications, the inability of any single storage structure to respond to these demands, and the slow decline of, disk access times (latency and seek) relative to the rapid increase in CPU speeds, memory size, and network bandwidth.We present a multi-structured file system designed for high bandwidth I/O and fast response. Our design is based on combining disk caching with three different file storage structures, each implemented on an independent and isolated disk array. Each storage structure is designed to optimize a different set of file system access characteristics such as cache writes, directory searches, file attribute requests or large sequential reads/writes.As part of our study, we analyze the performance of an existing file system using trace data from UNIX disk I/O-intensive workloads. Using trace driven simulations, we show how performance is improved by using separate storage structures as implemented by a multi-structured file system.
- 1.Britton, D., and Gray, J., "Disk Shadowing", Proceedings of the 14th Very Large Data Base Conference, September 1988, pp. 331-338. Google ScholarDigital Library
- 2.Chefiton, D. and Finalayson, R., "Log Files: An Extended File Service Exploiting Write-Once Storage.", Proceedings Eleventh Symposium on Operating Systems Principles, November 1987, pp. 139-148. Google ScholarDigital Library
- 3.Ferrari, D., "Computer Systems Performance Evaluation.", Prentice Hall, 1981.Google Scholar
- 4.Floyd, R., and Ellis, C., "Directory Reference Patterns in Hierarchical File Systems", iEEE Transactions on Knowledge and Data Engineering, Volume 1, Number 2, June 1989, pp. 238-247. Google ScholarDigital Library
- 5.Fujitsu M2361A Mini-Disk Drive Customer Engineering Manual.Google Scholar
- 6.Gait, J., "The Optical File Cabinet: A Random- Access File System For Write-Once Optical Disks.", IEEE Computer, Volume 21, Number 6, June 1988, pp. 11-22 Google ScholarDigital Library
- 7.Garcia-Molina, H., and Salem, K., "Disk Striping.", International Conference on Data Engineering, February 1986, pp. 336-342. Google ScholarDigital Library
- 8.Hagmann, R., "Reimplementing the Cedar File System Using Logging and Group Commit.", Proceedings Eleventh Symposium on Operating Systems Principles, November 1987, pp. 155-162. Google ScholarDigital Library
- 9.Katz, R., Gibson, G., and Patterson, D., "Disk System Architectures for High Performance Computing." Technical Report, University of California, Berkeley, UCB/CSD 89/497, March 1989, pp. 1-39. Google ScholarDigital Library
- 10.Kim, M., "Synchronized Disk Interleaving.", IEEE Transactions on Computers, Volume C-35, Number 11, November 1986, pp. 978-988. Google ScholarDigital Library
- 11.Lazowska, E. et al., "File Access Performance of Diskless Workstations.", ACM Transaction on Computer Systems, Volume 4, Number 3, August 1986, pp. 238-268. Google ScholarDigital Library
- 12.Livny, M, Khoshafian, S., and Boral, H., "Multi- Disk Management Algorithms.", Proceedings Eleventh Symposium on Operating Systems Principles, November 1987, pp. 69-76. Google ScholarDigital Library
- 13.McKusick, M.K, et al., "A Fast File System for UNIX.", Transactions on Computer Systems, Volume 2, Number 3, August 1984, pp. 181-197. Google ScholarDigital Library
- 14.Mullender, S., and Tanenbaum, A., "Immediate Files", Software Practice and Experience, Volume 14, Number 4, April 1984, pp. 365-368. Google ScholarDigital Library
- 15.Nelson, M., "Physical Memory Management in a Network Operating System.", Ph.D. Thesis, University of California, Berkeley, November 1988. Google ScholarDigital Library
- 16.Nelson, M., Welch, B., and Ousterhout, J. "Caching in the Sprite Network File System,", ACM Transactions on Computer Systems, Volume 6, Number 1, February 1988, pp. 134-154. Google ScholarDigital Library
- 17.Ohta, M,, and Tezuka, H., "A Fast/tmp File System by Delay Mount Option", USEN1X. Summer '90 Technical Conference, June 1990, pp. 145-150.Google Scholar
- 18.Ousterhout, J., et al., "A Trace Driven Analysis of the Unix 4.2 BSD File System.", Proceedings Tenth Symposium on Operating Systems Principles, ~mber 1985, pp. 15-24. Google ScholarDigital Library
- 19.Ousterhout, J., and Douglis, F., "Beating the I/O Bottleneck: The Case for Log-Structured File Systems.'', Operating Systems Review, January 1989, pp. 11-27, Google ScholarDigital Library
- 20.Patterson, D., Gibson, G., and Katz, R., "A Case for Redundant Arrays of Inexpensive Disks (RAID).", ACM SIGMOD 88, June 1988, pp. 109-116. Google ScholarDigital Library
- 21.Powell, M., "The DEMOS File System.", Proceedings Sixth Symposium on Operating Systems Principles, November 1977, pp. 33-42. Google ScholarDigital Library
- 22.Reddy, A., and Banerjee, P., "Performance Evaluation of Multiple-Disk I/O Systems.", 1989 International Conference of Parallel Processing, June 1989, pp. 315-318.Google Scholar
- 23.Rosenblum, M,, and Ousterhout, J., "The LFS Storage Manager", USENIX-Summer '90 Technical Conference, June 1990, pp. 315-324.Google Scholar
- 24.Satyanarayanan, M., "A Study of File Sizes and Functional Lifetimes.", Procee&'ngs Eighth Symposium on Operating Systems Principles, December 1981, pp. 96-108. Google ScholarDigital Library
- 25.Smith, A., "Disk Cache - Miss Ratio Analysis and Design Considerations.", ACM Transactions on Computer Systems, Volume 3, Number 3, August 1985, pp. 161-203. Google ScholarDigital Library
- 26.Svobodova, L., "A Reliable Object-Oriented Data Repository for a Distributed Computer System.", Proceedings Eighth Symposium on Operating Systems Principles, December 1981, pp. 47-58. Google ScholarDigital Library
- 27.Van Renesse, R., Tanenbaum, A. S., and Wilschut, A., "The Design of a High-Performance File Server", IEEE Transactions on Knowledge and Data Engineering, Volume 1, Number 2, june 1989, pp. 22-27.Google Scholar
- 28.Yokoyama, S. and Yamada, S., "A Contiguous High Performance File System", EUUG Spring '89, April 1989, pp. 197-206.Google Scholar
Index Terms
- A high performance multi-structured file system design
Recommendations
A high performance multi-structured file system design
File system I/O is increasingly becoming a performance bottleneck in large distributed computer systems. This is due to the increased file I/O demands of new applications, the inability of any single storage structure to respond to these demands, and ...
The design and implementation of a log-structured file system
This paper presents a new technique for disk storage management called a log-structured file system. A log-structured file system writes all modifications to disk sequentially in a log-like structure, thereby speeding up both file writing and crash ...
Comments