THE ZEBRA STRIPED NETWORK FILE SYSTEM

Citation
Jh. Hartman et Jk. Ousterhout, THE ZEBRA STRIPED NETWORK FILE SYSTEM, ACM transactions on computer systems, 13(3), 1995, pp. 274-310
Citations number
39
Categorie Soggetti
Computer Sciences","Computer Science Theory & Methods
ISSN journal
07342071
Volume
13
Issue
3
Year of publication
1995
Pages
274 - 310
Database
ISI
SICI code
0734-2071(1995)13:3<274:TZSNFS>2.0.ZU;2-R
Abstract
Zebra is a network file system that increases throughput by striping t he file data across multiple servers. Rather than striping each file s eparately, Zebra forms all the new data from each client into a single stream, which it then stripes using an approach similar to a log-stru ctured file system. This provides high performance for writes of small files as well as for reads and writes of large files. Zebra also writ es parity information in each stripe in the style of RAID disk arrays; this increases storage costs slightly, but allows the system to conti nue operation while a single storage server is unavailable. A prototyp e implementation of Zebra, built in the Sprite operating system, provi des 4-5 times the throughput of the standard Sprite file system or NFS for large files and a 15-300% improvement for writing small files.