Resolving File Conflicts in the FICUS File System
Peter Reiher, John Heidemann, David Ratner, Greg Skinner, Gerald Popek
Department of Computer Science
University of California, Los Angeles
Abstract
Ficus is a flexible replication facility with optimistic concurrency
control designed to span a wide range of scales and network
environments. Optimistic concurrency control provides rapid local
access and high availability of files for update in the face of
disconnection, at the cost of occasional conflicts that are only
discovered when the system is reconnected. Ficus reliably detects all
possible conflicts. Many conflicts can be automatically resolved by
recognizing the file type and understanding the file's semantics.
This paper describes experiences with conflicts and automatic conflict
resolution in Ficus. It presents data on the frequency and character
of conflicts in our environment. This paper also describes how
semantically knowledgeable resolvers are designed and implemented, and
discusses our experiences with their strengths and limitations. We
conclude from our experience that optimistic concurrency works well in
at least one realistic environment, conflicts are rare, and a large
proportion of those conflicts that do occur can be automatically
solved without human intervention.
Download the full text of this paper in
ASCII (64,419 bytes) and
POSTSCRIPT (131,990 bytes) form.
To Become a USENIX Member, please see our
Membership Information.