home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!sun-barr!ames!agate!darkstar.UCSC.EDU!osr
- From: vander@flab.fujitsu.co.jp (Mike van der Velden)
- Newsgroups: comp.os.research
- Subject: Summary: File Replication
- Date: 29 Aug 1992 05:02:55 GMT
- Organization: Fujitsu Laboratories Ltd., Kawasaki, Japan
- Lines: 159
- Approved: comp-os-research@ftp.cse.ucsc.edu
- Message-ID: <17n0dvINN7b0@darkstar.UCSC.EDU>
- NNTP-Posting-Host: ftp.cse.ucsc.edu
- Originator: osr@ftp
-
- Earlier this month, I wrote:
-
- > Does anyone have or know of some documentation about file replication,
- > especially within a Unix-like environment?
- >
- > I will summarize to the net if there is interest.
- >
- > Thanks in advance for your time.
-
- And now, thr promised summary. Thanks to the following people for
- taking the time to respond.
-
- page@ficus.cs.ucla.edu (Thomas Page)
- johnh@ficus.cs.ucla.edu (John Heidemann)
- paris@cs.uh.edu (Jehan-Francois Paris)
- mootaz@cs.rice.edu (Elmootazbellah Nabil Elnozahy)
- khattra@sfu.ca (Taj Khattra)
- Bob Gruber <gruber@thor.lcs.mit.edu>
- akyurek@cs.umd.edu (Sedat Akyurek)
- Jonathan Spencer <J.M.Spencer@newcastle.ac.uk>
- esprit@netcom.com (Alan F. Perry)
-
-
- --------------------------------------------------------------------------
- From: page@ficus.cs.ucla.edu (Thomas Page)
-
- > Our group at UCLA has produced a replicated file system, called Ficus,
- > for Sun OS. If interested, you can find papers about it available via
- > ftp from ftp.cs.ucla.edu in the director pub/ficus. The tech reports
- > from 1991 or 1992 are the most relevent.
-
-
- --------------------------------------------------------------------------
- From: johnh@ficus.cs.ucla.edu (John Heidemann)
-
- > You might be interested in the Ficus research here at UCLA.
- > We provide optimistic file replication under SunOS.
- > Technical reports are available by anonymous ftp from ftp.cs.ucla.edu.
- >
- > If you wish to provide replication by hand in a stock Unix system
- > you should check out rdist(1).
-
-
- --------------------------------------------------------------------------
- From: paris@cs.uh.edu (Jehan-Francois Paris)
-
- > There are for the moment much more proposals of replicated file systems
- > than actual implementations. The last Symposium on Operating Systems
- > Principles had two papers on file replication. The first paper,
- > "Disconnected Operation in the Coda File System, " by J. J. Kistler and
- > M. Satyanarayanan describes a system implementing an optimistic
- > replication protocol, i.e., a system trading the risk of inconsistent
- > updates for higher data availabilities. The second paper, "Replication
- > in the Harp File System," describes a replicated file system based
- > on the pessimistic approach.
- >
- > Most papers on file replication focus on the design of the protocols
- > used for guaranteeing the consistency of the replicated data. I do
- > not know of any recent bibliography on the topic but wrote last year
- > what I think is a decent survey of data replication for the Encyclopedia
- > of Microcomputing. Please let me know if you cannot get a copy of it.
-
-
- --------------------------------------------------------------------------
- From: mootaz@cs.rice.edu (Elmootazbellah Nabil Elnozahy)
-
- > Look at
- > "A Highly Available Network File Server", by A. Bhide, E. Elnozahy,
- > and S. Morgan, Winter Usenix 91.
- >
- > This shows that using replication is not necessary to obtain
- > high availability.
-
-
- --------------------------------------------------------------------------
- From: khattra@sfu.ca (Taj Khattra)
-
- > A nice survey article is
- >
- > Tait, C.D., _Techniques for Building Highly Available Distributed
- > File Systems_, CUCS-497-89.
- >
- > It can be anonymous ftp'ed from cs.columbia.edu in
- > /pub/reports/reports-1989/cucs-497-89.ps.Z
-
-
- --------------------------------------------------------------------------
- From: Bob Gruber <gruber@thor.lcs.mit.edu>
-
- > Well, you could start with our paper on a replicated NFS server:
- >
- > "Replication in the Harp File System"
- > Liskov, Ghemaway, Gruber, Johnson, Shrira, and Williams
- > In the Proceedings of the 13th ACM Symposium on Operating
- > System Principles (SOSP), October 1991.
- >
- > This paper references other replicated file systems; see
- > the related work section.
-
-
- --------------------------------------------------------------------------
- From: akyurek@cs.umd.edu (Sedat Akyurek)
-
- > I don't know if you are looking for software/systems that do replication or
- > works/papers on this topic but here are some references to some
- > papers on this topic :
- >
- > Liskov, Barbara, et. al., ``Replication in the Harp File System,''
- > Proceedings of the 13th ACM Symposium on Operating System Principles,
- > October, 1991.
- >
- > Lo, Sai-Lai, ``Ivy: A Study on Replicating Data for Performance
- > Improvement,'' Technical Report, Hewlett-Packard Laboratories, Concurrent
- > Computing Department, HPL-CSP-90-48, December 1990.
- >
- > Page, Thomas W., et. al., ``Management of Replicated Volume Location in
- > the Ficus Replicated File System,'' USENIX Conference Proceedings, Summer
- > 1991.
- >
- > I am working on replication in block level (on the same disk) for reducing
- > seek latency (and thus for reducing disk request response times).
- > An introduction to my work is in :
- >
- > Akyurek, Sedat, Kenneth Salem, ``Placing Replicated Data to Reduce Seek
- > Delays,'' USENIX File Systems Workshop Proceedings, May 1992.
-
-
- --------------------------------------------------------------------------
- From: Jonathan Spencer <J.M.Spencer@newcastle.ac.uk>
-
- > Try: Mark Little's PhD thesis about replication in object oriented
- > distributed systems, dated Sep 1991. Should be available through
- > inter-library loans. (Oh, he did it here :-)
- >
- > Mail : Computing Lab., University of Newcastle-upon-Tyne, NE1 7RU, England
- > ARPA : J.M.Spencer%newcastle.ac.uk@cs.ucl.ac.uk
- > JANET : J.M.Spencer@newcastle.ac.uk
- > UUCP : !uknet!newcastle.ac.uk!J.M.Spencer Phone : +91 222 8229
-
-
- --------------------------------------------------------------------------
- From: esprit@netcom.com (Alan F. Perry)
-
- > I am sure that you know that file replication is done in Locus.
-
-
- --
- Mike van der Velden
- Distributed Systems Laboratory, Fujitsu Laboratories Ltd., Kawasaki, Japan
- e-mail: vander@tokyo.hi.flab.fujitsu.co.jp
-
- ps. If you send me a picture postcard Mike van der Velden
- from your area to the address at Fujitsu Unoki Dormitory, room 227
- right, I'll send you one from Tokyo. 2-17-18 Unoki, Ota-ku, Tokyo 146
- Postcards must be received by Japan
- November 15, 1992
-
-
-
-