Talk:Network file system

"DFS" disambiguation

DFS does also mean Dynamic Frequency Selection. I suggest to create a disambiguation page here. Can someone do this for me? I have no Idea how to create something like this...

I think a catalog like Comparison of file systems would be handy.

There is no such thing as DFS

says 68.231.243.104 who has removed all the content 4 or 5 times in the past week.

68.231.243.104, instead of removing the content, why not replace it with something explaining why there is not such thing as a DFS. Or better yet, leave the content alone and discuss it on the talk page. You your creating making a nuisance of yourself by not talking. Tom Brown 22:41, 22 Apr 2005 (UTC)

Distributed, ok, but, that is not distributed

A distributed file system is a network file system that works distributed accross many servers.

That is, you create a share that contains files from more than one server, and usually supporting replication and offline access.

The article is talking about plain network file systems.

Can someone correct that?

—Claunia 21:34, 2 September 2005 (UTC)[reply]

Not sure what the article looked like when you wrote this comment, but at present the article has distributed filesystems, point to point server->client filesystems, distributed databases, point to point databases, and other miscellany all piled together. I don't even know where to begin cleaning this mess up. 64.142.12.203 18:46, 17 January 2007 (UTC)[reply]

I started the cleanup with moving "distributed data store" to a separate article, and while at it, split the list into two categories (client-server file system vs distributed file system) to my understanding of these.

Next, we should agree on a clear set of definitions of the terms "distributed file system", "network file system", "cluster file system" and other related terms. And more importantly, whether concept one is a subset of another or not.

I would not propose a definition at this point, but to me, "distributed file system" seems to imply that the file system objects are being manipulated and controlled on different computers – definitely not that there's a single master that would take care of all this (the latter is the case with client-server file systems). Are there any file systems satisfying this criteria that aren't cluster file systems? -- intgr 23:26, 17 January 2007 (UTC)[reply]

Distributed file systems are already listed at List_of_file_systems#Distributed_file_systems so I think we should just reference the better list rather than have a lesser list here. Angelo 20:00, 8 February 2007 (UTC)[reply]

some other dfs

there are

(H)DFS: Hadoop's Distributed File System ( http://wiki.apache.org/lucene-hadoop/DFS , http://wiki.apache.org/lucene-hadoop/HadoopOverview )

and

GFS: GoogleFileSystem ( http://labs.google.com/papers/gfs.html )

and

NDFS: Nutch Distributed File System ( http://wiki.apache.org/nutch/NutchDistributedFileSystem )

too

SVK as a distributed file system ?

SVK is an SCM. I can't see the point to list it in the distributed file system ? —Preceding unsigned comment added by AlexandreDulaunoy (talk • contribs)

Yes, doesn't make sense to me, although the definition of a "distributed file system" is rather vague. Removed -- intgr 15:40, 28 December 2006 (UTC)[reply]

Distributed, networked and cluster file systems

As far as I can understand, the article's definition seems to say that a "distributed file system" is a file system that allows access to individual files over the network.

However, some cluster file systems also call themselves "distributed file systems". These file systems run on each cluster node and are "distributed" in the sense that they access the same disk storage backend (block device) over the network and keep all other nodes in sync with updates; this approach is, however, very different from usual networked file systems that allow access directly to individual files and directories over the network.

I think the article should more clearly state what the different terms mean. -- intgr 16:00, 28 December 2006 (UTC)[reply]

As per the confusion/comments in above posts, I am going ahead and moving this article to "Network file system" -- intgr 16:22, 28 December 2006 (UTC)[reply]

Cannot move, a redirect is in the way. Will propose to requested moves instead. -- intgr 22:15, 28 December 2006 (UTC)[reply]

The following discussion is an archived debate of the proposal. Please do not modify it. Subsequent comments should be made in a new section on the talk page. No further edits should be made to this section.

The result of the debate was Move Network File System → Network File System (Sun) and then distributed file system → network file system. The only opposition was based on name clash and this is a good counterproposal. If another disambiguated title gains consensus, it's unlikely to be blocked anyway so feel free to handle that here. —Wknight94 (talk) 13:07, 6 January 2007 (UTC)[reply]

Requested move

distributed file system → network file system — The article really talks about simply network file systems – that is, file systems that are accessed over a network. Cluster file systems are often called "distributed file systems" and they are conceptually quite distinct from network file systems. intgr 22:21, 28 December 2006 (UTC)[reply]

Survey

Add # '''Support''' or # '''Oppose''' on a new line in the appropriate section followed by a brief explanation, then sign your opinion using ~~~~.

Survey - Support votes

Survey - Oppose votes

Oppose. Network File System is trademarked and network file system redirects to the article about it, as it should. Trying to distinguish based on upper/lower case is discouraged. --Serge 19:00, 2 January 2007 (UTC)[reply]
Do you have any better suggestions how to resolve the situation? -- intgr 04:49, 3 January 2007 (UTC)[reply]
Perhaps Network File System could be moved to Network File System (protocol), to further distinguish it from network file system? -GTBacchus^(talk) 02:17, 6 January 2007 (UTC)[reply]
Yeah, adding a disambiguation title is a good idea. Perhaps "protocol" is not the best one since there are different incompatible protocols labeled as NFS. Maybe "Network File System (Sun)" -- intgr 06:51, 6 January 2007 (UTC)[reply]

Discussion

Add any additional comments:

I tend to agree on the "unclear" definition regarding distributed file system. The only issue (IMHO) I see is the existing confusion between NFS and using the generic name network file system. For example (http://foldoc.org/index.cgi?query=network+file+system&action=Search) FOLDOC is referencing only to NFS (the one from Sun) and not the generic term. I think the renaming won't directly solve the problem. It could be better to improve the current definition of distributed file system. Just my point of view. AlexandreDulaunoy 10:11, 29 December 2006 (UTC)[reply]

The Network File System article already has a {{for}} template in the beginning; do you think having these on both articles wouldn't be clear enough? -- intgr 20:25, 29 December 2006 (UTC)[reply]

The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made in a new section on this talk page. No further edits should be made to this section.

If people want to insist on Network_File_system as a generic term ...

By the way file, "remote file access protocol" is more precise.

At least have a disambiguation.

Also Adding " S(un)" as a suffix is incorrect as the IETF trust owns NFS.

What's next, making "File Transfer Protocol" and "Secure Shell" generic terms? Sheesh. —The preceding unsigned comment was added by Mre5765 (talk • contribs) 19:24, 31 January 2007 (UTC).[reply]

The following comment was added by this user after the nomination had been closed (diff):

Oppose. "Network FIle System (Sun)" is wrong. How did this happen with no approve votes and all oppose votes. This is history.

-- Mike Eisler, co-author RFC 3530.

First, take a deep breath and calm down; you appear to be taking this issue too personally. :)

They are commonly referred to as "file systems" and they work over a network. How exactly is using "network file system" not appropriate for a generic title? I think it's much more appropriate than "distributed file system". Evidently, I am not the only person who calls these protocols "network file system" — they're also often labeled such in academics. For example, see a Google Scholar search for this phrase: [1]

As for disambiguation, there already is a "For the Sun protocol ..." link at the beginning of the article. Separate disambiguation pages are usually not created when there are only two subjects in the same field, and one can be considered primary (which in this case would be the generic meaning).

Obviously you have the right to disagree with the moves, and I am open to discussion. The closed move nomination is not set in stone and can be fixed by moving again.

Note that nomination itself also counts as a "for" argument, and the only "oppose" vote was based on disambiguating the articles by case, which was solved with adding a disambiguation suffix (and yes, I am aware of your objections to using that). Nominations are discussions and not votes, although they are often misinterpreted as such (see Wikipedia:Polling is not a substitute for discussion)

Also, you should not change nominations after they have been closed, as the red and bold text "Please do not modify it" exclaims — I will now proceed to remove your comment. Please do not interpret this as an attack. -- intgr 20:57, 31 January 2007 (UTC)[reply]

Dubious link to ETUfs?

Something called "ETUfs" is mentioned in the list of filesystems. It (so far) links to nothing on Wikipedia, and a search with Google returns no hits. It's quite strange that any filesystem wouldn't be mentioned on webpages, so I'm wondering, does it even exist? If it does, some more words on where it comes from / what its specifics are / a link to more information, seem in order.— Preceding unsigned comment added by 193.190.253.146 (talk • contribs) 08:47, 9 June 2007