Conquest File System

NSF grant CCR-0098363

Charles Weddle, Mathew Oldham, Florida State University
RuGang Xu, UCLA, The Laboratory for Advanced Systems Research
An-I Andy Wang , Florida State University
Geoffrey H. Kuenning, Harvey Mudd College
Peter Reiher and Gerald J. Popek, UCLA, The Laboratory for Advanced Systems Research

The rapidly declining cost of persistent RAM technologies prompts the question of when, not whether, such memory will become the preferred storage medium for many computers.  In anticipation of that development, we present Conquest, a file system that provides a transition from disk to persistent RAM as the primary storage medium. Conquest is incrementally deployable and realizes most of the benefits of persistent RAM as it becomes cheaply abundant.  As of October 2001, Conquest can be used effectively for a hardware cost of below $200.

We compare Conquest's performance to ext2, reiserfs, SGI XFS, and ramfs, using popular benchmarks.  Our measurements show that Conquest incurs little overhead compared to ramfs.  Compared to disk-based file systems, Conquest achieves 24% to 1900% faster performance for working sets that fit in memory, and 43% to 96% faster performance with working sets larger than the memory size.

Conquest Architecture

Conquest uses memory to store all metadata, small files (currently based on a size threshold), executables, and shared libraries, leaving only the content of large files on disk.  All accesses to in-core data and metadata incur no data duplication or disk-related overhead, and executions are in-place.  For the large-file-only disk storage, we use a larger access granularity to reduce the seek-time overhead.  Because most accesses to large files are sequential, we can relax many historical disk design constraints, such as complex layout heuristics intended to reduce fragmentation or average seek times.
 
 

Publications

Wang06
An-I Andy Wang, Geoff Kuenning, Peter Reiher, and Gerald Popek. The Conquest File System: Better Performance Through a Disk/Persistent-RAM Hybrid Design. ACM Transactions on Storage (TOS), 2(3), pp. 309-348, August 2006. [24% acceptance rate for the original USENIX ATC conference paper]

Wang03b
An-I Andy Wang, Geoffrey H. Kuenning, Peter Reiher, Gerald J. Popek. The Effects of Memory-Rich Environments on File System Microbenchmarks. Proceedings of the 2003 International Symposium on Performance Evaluation of Computer and Telecommunication Systems (SPECTS), Montreal, July 2003.<http://www.cs.fsu.edu/~awang/papers/spects2003a.html>

Wang03a
An-I Andy Wang. The Conquest File System: A Disk/Persistent-RAM Hybrid Design for Better Performance and Simpler Data Paths. Ph.D. Dissertation. Computer Science Department, University of California, Los Angeles, March 2003.<http://www.cs.fsu.edu/~awang/papers/dissertation2003a.html>

Wang02b
An-I Andy Wang, Geoffrey H. Kuenning, Peter Reiher, Gerald J. Popek. Conquest: Better Performance Through a Disk/Persistent-RAM Hybrid File System. Proceedings of the 2002 USENIX Annual Technical Conference, Monterey, June 2002. <http://www.cs.fsu.edu/~awang/papers/usenix2002a.html>

Wang02a
An-I Andy Wang, Geoffrey H. Kuenning, Peter Reiher, Gerald J. Popek. Work-in-Progress Report: Conquest: Better Performance Through a Disk/Persistent-RAM Hybrid File System. On-Line Proceedings of the First USENIX Conference on File and Storage Technologies (FAST), Monterey, January 2002.<http://www.cs.fsu.edu/~awang/papers/fast2002a.html>
 
Wang01a
An-I Andy Wang, Geoffrey H. Kuenning, Peter Reiher, Gerald J. Popek. Position Summary: The Conquest File System--Life after Disks. Proceedings of the 8th IEEE Workshop on Hot Topics in Operating Systems (HotOS VIII), Schloss Elmau, Germany, May 2001.<http://www.cs.fsu.edu/~awang/papers/hotos2001.html>

Presentations


RuGang Xu. Conquest: Combining Battery-Backed RAM and Threshold-Based Storage Scheme to Conserve Power.  Presented at the 19th Symposium on Operating Systems Principles (SOSP), New York, October 2003.
 

An-I Andy Wang. Conquest: Preparing for Life After Disks.  Presented at the Computer Science Colloquium, Florida State University, Tallahassee, October 2003.
 

Geoffrey H. Kuenning. The Disk is Dead! Long Live the Disk!.  Presented at the Computer Science Colloquium, Harvey Mudd College, Claremont, April 2003.
 

An-I Andy Wang. Conquest:  Preparing for Life After Disks.  Presented at the University of California, Los Angeles, Advanced Operating Systems Lecture, Los Angeles, October 2002.
 

An-I Andy Wang. Conquest:  Better Performance Through a Disk/Persistent-RAM Hybrid File System.  Presented at the 2002 USENIX Annual Technical Conference, Monterey, June 2002.
 

An-I Andy Wang. The Conquest File System:  Life After Disks.  Poster presented at the UCLA Computer Science Department Annual Research Review, UCLA, Los Angeles, April 2002.
 

An-I Andy Wang. Conquest:  Better Performance Through a Disk/Persistent-RAM Hybrid File System.  Presented at the First USENIX Conference on File and Storage Technologies (FAST), Monterey, January 2002.
 

An-I Andy Wang. Conquest:  RAM as Storage; Disks as Tapes.  Presented at the University of California, Los Angeles, Advanced Operating Systems Lecture, Los Angeles, November 2001.
 

An-I Andy Wang.  The Conquest File System:  Life after Disks.  Poster presented at the UCLA Computer Science Department Annual Research Review, UCLA, Los Angeles, April 2001.
 

An-I Andy Wang. Integration of Memory and File System Services via Persistent RAM.  Presented at the Computer Science Colloquium, Harvey Mudd College, Claremont, October 1999.

Press Coverage


Conquest FS: The Disk is Dead, Slashdot, April 21 2003.
 

Eugenia Loli-Queru. Learning About the Conquest File System, OS News, April 21 2003.
 

Shawn. Not Your Father's File System, hartys.net::the blog, April 21 2003.
 

Baltazar H.  DAFS, Conquest May Pave Way for Future File Systems, eWeek, Ziff Davis Media, June 17 2002.
 

Conquest File System.  LisoLeg: Linux Source Learning Group, 2000.

awang@cs.fsu.edu

Last modified:  November 19, 2001