[264] in Commercialization & Privatization of the Internet
Alternatives to NFS (Was: Re: Should the NREN be funded?)
daemon@ATHENA.MIT.EDU (Manavendra K. Thakur)
Sun Mar 3 04:50:46 1991
To: stev@ftp.com (stev knowles)
Cc: com-priv@psi.com
In-Reply-To: Your message of Sat, 02 Mar 91 23:14:17 -0500.
Date: Sun, 03 Mar 91 04:21:32 EST
From: "Manavendra K. Thakur" <thakur@zerkalo.harvard.edu>
>>>>> On Sat, 2 Mar 91 23:14:17 -0500, stev@ftp.com (stev knowles) said:
>> I agree with Steve that more funding for University research is
>> needed, but I'm not sure I agree that NFS is non-changable because
>> it is "owned" by Sun. The reason it's hard to change is because so
>> many things out there use it and depend on it being the way it is
>> for interoperability. So many non-Sun vendors have NFS in their
>> OS's now (or have NFS available via third-party vendors) that making
>> any major changes just wouldn't work.
> the common answer i use to this question has to do with checksums.
> Sun refuses to add checksums to NFS, for "speed considerations".
> this causes other vendors, who simply repackage the reference port
> of NFS, to resist changing the code.
Apparently Sun already has added checksums to NFS - I'm told Sun's C2
security package has that feature. I'm also told that the C2 package
doesn't work all that well or smoothly, so perhaps in practice that is
why you say Sun refuses to add checksums to NFS. Also, I don't know
whether other vendors provide checksum capabilities in their
implementations of NFS.
But I hasten to point out that adding checksums is hardly enough to
correct the deficiencies of NFS! Largely because NFS is both
stateless and synchronous, NFS is simply too slow and inefficient for
building large-scale filesystems that might span nations or
continents. Nor is there any mechanism in NFS to create and enforce a
global namespace (except by self-imposed discipline, which is doable
but makes NFS more difficult to administer). For these and other
reasons, NFS is not scalable in any meaningful sense of the word.
Research *has* been done into alternatives, and while it's clear that
NFS is the de facto standard today, much better technologies exist and
are in production use already.
For example, research into remote file sharing has been done at
Carnegie-Mellon University since at least 1986 and probably well
before that. Out of that university research - and this a textbook
example of the type of funding that NREN should provide to
universities - has emerged the Andrew File System (AFS), which
corrects many of the deficiencies of NFS. AFS is, to my mind, two
generations ahead of NFS, technologically speaking.
Since 1989, AFS has been a commercial product available from Transarc
Corp. in Pittsburgh, and it now runs on all the major workstation
platforms (including Suns, Decs, IBM, HP, NeXt, etc). Transarc also
has developed an NFS-to-AFS translator that allows AFS server machines
to speak NFS to those machines that cannot or do not support AFS.
Also, the Open Software Foundation has incorporated AFS into its DCE
(Distributed Computing Environment) suite.
Similarly, AT&T has something called RFS, which is available on Suns
as of SunOS 4.1. I haven't had time to play with it (as I've done
with AFS), so I don't know much about the capabilities of RFS. But it
is there and available for those who want to explore it.
Finally, BSD 4.4 has a publicly available (not public domain)
implementation of NFS. You could always ftp this code and hack on it
as much as you like.
So the point of this note is that as far as Internet standards are
concerned (de facto or otherwise), none of us are really at Sun's
mercy when it comes to accessing files remotely and transparently
across a network. Don't just blindly assume you're stuck with your
current technologies if the vendor refuses to change something. There
are almost always alternatives, and good ones at that, which are well
worth exploring.
The backers of NREN would do well to look at the history of AFS as it
developed out of university research and migrated to a commercial
product. Presenting AFS as a case study to Congress might help
persuade that august body to fund similar endeavors in the future to
help fulfill both the research and educational missions of NREN.
Manavendra K. Thakur Internet: thakur@zerkalo.harvard.edu
Systems Programmer, High Energy Division BITNET: thakur@cfa.BITNET
Harvard-Smithsonian Center for DECNET: CFA::thakur
Astrophysics UUCP: ...!uunet!mit-eddie!thakur
P.S. Disclaimer - I don't work for nor do I get paid by Transarc, CMU,
AT&T, Berkeley, or Sun.
