DCE Distributed File System
Encyclopedia
The DCE Distributed File System (DCE/DFS) is the remote file access protocol used with the Distributed Computing Environment
. It was based on the AFS Version 3.0 protocol that was developed commercially by Transarc
Corporation. AFS Version 3.0 was in turn based on the Andrew File System
Version 2.0 protocol (also used by the Coda disconnected file system) originally developed at Carnegie Mellon University
.
DCE/DFS consisted of multiple cooperative components that provided a network file system with strong file system semantics, attempting to mimic the behavior of POSIX
local file systems while taking advantage of performance optimizations when possible. A DCE/DFS client system utilized a locally managed cache that would contain copies (or regions) of the original file. The client system would coordinate with a server system where the original copy of the file was stored to ensure that multiple clients accessing the same file would re-fetch a cached copy of the file data when the original file had changed.
The advantage of this approach is that it provided very good performance even over slow network connections because most of the file access was actually done to the local cached regions of the file. If the server failed, the client could continue making changes to the file locally, storing it back to the server when it became available again.
DCE/DFS also divorced the concept of logical units of management (Fileset
s) from the underlying volume on which the fileset was stored. In doing this it allowed administrative control of the location for the fileset in a manner that was transparent to the end user. To support this and other advanced DCE/DFS features, a local journaling file system (DCE/LFS also known as Episode
) was developed to provide the full range of support options.
IBM has not maintained it since 2005: http://www-306.ibm.com/software/stormgmt/dfs/
IBM was working on a replacement for DCE/DFS called ADFS (Advanced Distributed File System). One major goal of this project was to decouple DFS from the complexities of DCE's cell directory services (CDS) and security services (secd). Another key feature would have been the elimination of enctype limitations associated with DCE/RPC. No public mention of this effort has been made since 2005, leading many to believe the project has been killed.
Distributed Computing Environment
The Distributed Computing Environment is a software system developed in the early 1990s by a consortium that included Apollo Computer , IBM, Digital Equipment Corporation, and others. The DCE supplies a framework and toolkit for developing client/server applications...
. It was based on the AFS Version 3.0 protocol that was developed commercially by Transarc
Transarc
Transarc Corporation was a private Pittsburgh-based software company founded in 1989 by Jeffrey Eppinger, Michael Kazar, Alfred Spector, and Dean Thompson of Carnegie Mellon University...
Corporation. AFS Version 3.0 was in turn based on the Andrew File System
Andrew file system
The Andrew File System is a distributed networked file system which uses a set of trusted servers to present a homogeneous, location-transparent file name space to all the client workstations. It was developed by Carnegie Mellon University as part of the Andrew Project. It is named after Andrew...
Version 2.0 protocol (also used by the Coda disconnected file system) originally developed at Carnegie Mellon University
Carnegie Mellon University
Carnegie Mellon University is a private research university in Pittsburgh, Pennsylvania, United States....
.
DCE/DFS consisted of multiple cooperative components that provided a network file system with strong file system semantics, attempting to mimic the behavior of POSIX
POSIX
POSIX , an acronym for "Portable Operating System Interface", is a family of standards specified by the IEEE for maintaining compatibility between operating systems...
local file systems while taking advantage of performance optimizations when possible. A DCE/DFS client system utilized a locally managed cache that would contain copies (or regions) of the original file. The client system would coordinate with a server system where the original copy of the file was stored to ensure that multiple clients accessing the same file would re-fetch a cached copy of the file data when the original file had changed.
The advantage of this approach is that it provided very good performance even over slow network connections because most of the file access was actually done to the local cached regions of the file. If the server failed, the client could continue making changes to the file locally, storing it back to the server when it became available again.
DCE/DFS also divorced the concept of logical units of management (Fileset
Fileset
Fileset has several meanings, depending on the context:In the AIX operating system it is the smallest individually installable unit ....
s) from the underlying volume on which the fileset was stored. In doing this it allowed administrative control of the location for the fileset in a manner that was transparent to the end user. To support this and other advanced DCE/DFS features, a local journaling file system (DCE/LFS also known as Episode
Episode filesystem
Episode is a POSIX compliant file system most commonly known for its use in DCE/DFS file servers. It was designed to achieve the goals of portability, scaling, and to make more efficient use of available system bandwidth...
) was developed to provide the full range of support options.
IBM has not maintained it since 2005: http://www-306.ibm.com/software/stormgmt/dfs/
IBM was working on a replacement for DCE/DFS called ADFS (Advanced Distributed File System). One major goal of this project was to decouple DFS from the complexities of DCE's cell directory services (CDS) and security services (secd). Another key feature would have been the elimination of enctype limitations associated with DCE/RPC. No public mention of this effort has been made since 2005, leading many to believe the project has been killed.