Fstress: A Flexible Network File Service Benchmark

Benchmarks provide repeatable workloads for testing and comparison. An ideal file service benchmark would be easy to use, be highly configurable to emulate arbitrary workloads, and scale to extreme load levels and data set sizes. This paper introduces Fstress, a synthetic, flexible, self-scaling file service benchmark. We identify and support twelve workload-description parameters, enabling a wide range of workloads to evaluate file service scalability, sizing, configuration, and other factors. We demonstrate the generality of our approach by using these parameters to express seven popular existing workloads that exercise different aspects of a file service. As workload patterns continue to evolve, it is straightforward to capture and emulate their essential characteristics.

[1]  R. S. Fabry,et al.  A fast file system for UNIX , 1984, TOCS.

[2]  Mary Baker,et al.  Measurements of a distributed file system , 1991, SOSP '91.

[3]  Bruce E. Keith,et al.  LADDIS: The Next Generation in NFS File Server Benchmarking , 1993, USENIX Summer.

[4]  James Lau,et al.  File System Design for an NFS File Server Appliance , 1994, USENIX Winter.

[5]  Michael Dahlin,et al.  A quantitative analysis of cache policies for scalable network file systems , 1994, SIGMETRICS.

[6]  David A. Patterson,et al.  A new approach to I/O performance evaluation: self-scaling I/O benchmarks, predicted I/O performance , 1994, TOCS.

[7]  Maria Ebling,et al.  SynRGen: an extensible file reference generator , 1994, SIGMETRICS.

[8]  Martin F. Arlitt,et al.  Web server workload characterization: the search for invariants , 1996, SIGMETRICS '96.

[9]  Michael J. Feeley,et al.  The Measured Access Characteristics of World-Wide-Web Client Proxy Caches , 1997, USENIX Symposium on Internet Technologies and Systems.

[10]  Jeffrey Katcher,et al.  PostMark: A New File System Benchmark , 1997 .

[11]  Peter Druschel,et al.  A Scalable and Explicit Event Delivery Mechanism for UNIX , 1999, USENIX Annual Technical Conference, General Track.

[12]  B. Bershad,et al.  Manageability, availability and performance in Porcupine: a highly scalable, cluster-based mail service , 1999, SOSP.

[13]  Evangelos P. Markatos,et al.  Secondary Storage Management for Web Proxies , 1999, USENIX Symposium on Internet Technologies and Systems.

[14]  Jeffrey C. Mogul,et al.  Brittle metrics in operating systems research , 1999, Proceedings of the Seventh Workshop on Hot Topics in Operating Systems.

[15]  Ian W. Marshall,et al.  File popularity characterisation , 2000, PERV.

[16]  Yale N. Patt,et al.  Soft updates: a solution to the metadata update problem in file systems , 2000 .

[17]  Thomas E. Anderson,et al.  A Comparison of File System Workloads , 2000, USENIX Annual Technical Conference, General Track.

[18]  R. Card,et al.  Design and Implementation of the Second Extended Filesystem , 2001 .

[19]  Stefan Saroiu,et al.  A Measurement Study of Peer-to-Peer File Sharing Systems , 2001 .

[20]  Margo Seltzer,et al.  Workload-specific file system benchmarks , 2001 .

[21]  Syam Gadde,et al.  The Trickle-Down Effect: Web Caching and Server Request Distribution , 2002, Comput. Commun..