Talk:Comparison of file synchronization software

Entries to add

 * AeroFS - business plan only, but superior product — Preceding unsigned comment added by 184.100.191.38 (talk) 06:26, 5 January 2018 (UTC)
 * cloudplan - P2P file sync & share — Preceding unsigned comment added by 93.202.193.27 (talk) 20:28, 19 July 2016 (UTC)
 * Windows Offline files
 * Dropbox —Preceding unsigned comment added by 130.15.3.146 (talk) 16:18, 4 May 2011 (UTC)
 * Google Drive — Preceding unsigned comment added by 77.241.45.51 (talk) 06:07, 2 September 2012 (UTC)
 * Puresync
 * Microsoft OneDrive -- Guntherpb4u (talk) 15:18, 1 June 2015 (UTC)
 * Memeo autosync
 * SyncBreeze
 * SparkleShare —Preceding unsigned comment added by 67.68.23.7 (talk) 02:58, 11 March 2011 (UTC)
 * QtdSync
 * cwRsync --DavidCary (talk) 05:24, 26 February 2014 (UTC)
 * Amazon Cloud Drive -79.177.122.137 (talk) 22:59, 2 May 2014 (UTC)
 * Seafile : open source, free/libre Dropbox clone for self-hosting. Server available for Linux and Windows. Clients available for Android, Linux, iOS, Mac, and Windows. --ChPietsch (talk) 15:02, 12 August 2014 (UTC)
 * ChronoSync — Preceding unsigned comment added by 24.227.118.86 (talk) 15:32, 13 August 2014 (UTC)
 * SureSync  — Preceding unsigned comment added by Dereitz (talk • contribs) 18:58, 11 September 2014 (UTC)
 * Syncany — Preceding unsigned comment added by 45.48.115.76 (talk) 01:28, 23 March 2015 (UTC)
 * Csync --31.30.84.130 (talk) 12:06, 3 November 2021 (UTC)
 * PeerWasp — Preceding unsigned comment added by 45.48.115.76 (talk) 01:32, 23 March 2015 (UTC)
 * JottaCloud Randal L. Schwartz (talk) 21:51, 30 June 2016 (UTC)
 * Unison File Synchonizer Ericding (talk) 13:32, 31 May 2017 (UTC) Has been added by Drpixie (talk) 11:19, 2 July 2021 (UTC)

unnecessary "cleanup"
Someone ( Hm2k ) has deleted CleanSync and other items in as of 21:06, 10 May 2010



-- —Preceding unsigned comment added by 158.195.191.96 (talk • contribs)

system requirements
It should be useful to list also system requirements of compared software, because it has impact on performance. —Preceding unsigned comment added by 158.195.191.96 (talk) 11:29, 6 June 2010 (UTC)

wrong license type
Some items requests license agreements other than listed in the table ! —Preceding unsigned comment added by 212.55.255.41 (talk) 10:17, 8 June 2010 (UTC)

Three tables?!
I really don't understand this format. Like people only did care about certain features if tool is proprietary. They should be compared altogether, with the Licence as a column.

Having two tables would be useful if we want to split: general information, more important features, less important features.

And a column for "Paid version" is almost a joke. If there is no paid version available, I guess the software has no interest at all. --LQST (talk) 15:52, 21 June 2010 (UTC)


 * I agree, the tables should be combined and its licensing model (open source and which license/closed source/proprietary/freeware/shareware/paid) should be noted. Syzygy84 (talk) 21:22, 21 May 2014 (UTC)

Missing explanation
There are some software not made to work in Linux enviroments. I created a hypothesis: can a Windows laptop access a Linux shared partition (using Samba), and sync it's content, thanks to a Windows aplication that doesn't work in Linux? Can this sort of solution work? Or is it necessary to have the aplication installed in both machines? The answer to this could be put in the reference item. —Preceding unsigned comment added by 189.26.169.85 (talk) 11:15, 31 January 2011 (UTC)

Ambiguity
Does anyone else find the columns "Windows", "Mac OS X", "Linux", and "Other OS" ambiguous? For example, does it mean the application runs on those operating system platforms, or does it mean it just supports synchronizing with their filesystems?

Also, I don't understand why the commercial comparison chart has columns for both "Prior file versions, revision control" and "Restore replaced/deleted files from old versions." Isn't that the same feature? Even more confusing is that almost every piece of software says it can do revision control, but it also says it *can't* restore old files. I don't understand how that can be... isn't the whole point of revision control to be able to restore older versions?** [http://en.wikiquote.org/wiki/Zoolander#Mugatu They're the same thing! Doesn't anybody notice this? I feel like I'm taking crazy pills!] Jmontee (talk) 22:57, 14 October 2011 (UTC)

Reviewing the article history shows user "Alishoki" created this column "Restore replaced/deleted files from old versions." (including a period at the end of the column header), while indicating that only BestSync possessed this feature among all of the software shown. This suggests the intentions of this modification was in the interests of promoting a feature of BestSync. While in itself is perfectly acceptable, I draw upon personal experience with Risefly (author of BestSync) with which they seemed to possess less-than-desirable knowledge/quality control. I suspect this chart modification was performed by them without understanding that a column for the feature they wanted to illustrate already existed ("revision control"), and that it was incorrect to indicate that none of the other software possessed that feature. Additionally, Risefly is based in Osaka, and the username that made these modifications appears Japanese. For the sake of accuracy, I will manually revert these changes back and eliminate the redundant and inaccurate column for "Restore replaced/deleted files from old versions."
 * Update 17 October 2011 **

Table too wide
Someone on help desk just mentioned that the top table is too wide to view nicely. I am new, so I don't know what the WP policies/guidelines are for this.--Canoe1967 (talk) 05:35, 24 March 2012 (UTC)
 * Came here from the Help Desk post as well. I agree that table is too wide for most computer screens. Wikipedia:Manual of Style/Tables and WP:SPLITLIST seem to indicated that the way to handle this is removing some of the columns, since too much statistical data is against Wikipedia is not an indiscriminate collection of information policy. -- Uzma Gamal (talk) 15:37, 24 March 2012 (UTC)

Definitions
This article needs to include the definition of "portable" "scheduling" etc. for the top table. What do this terms mean, in this context? Is the data format portable? Or the software cross-platform? The software includes scheduling of batch file operations, or processes? What? 90.193.161.157 (talk) 10:38, 4 April 2012 (UTC)


 * I noticed confusion about "Detect rename, move" when testing (only) DirSync Pro and FreeFileSync for this aspect and corrected the table. So I started the section "Definitions".
 * UnTrueOrUnSimplified (talk) 20:44, 16 April 2013 (UTC)


 * A good start, but it's only a start. The table still has confusing columns such as "LAN Sync" and "Sync any directory" which can have multiple meanings. -79.177.122.137 (talk) 23:58, 6 May 2014 (UTC)

Call for user opinions about "Detect rename, move"
I want users opinions about the proposal to integrate the two columns "Detect rename", ".. move" to one with the name "Rename/move synced as such". I assume all programs detect added or deleted files, and with that, renamed and moved files. I see no point in reserving 2 columns for that definition. UnTrueOrUnSimplified (talk) 14:52, 18 April 2013 (UTC)
 * I agree! OGivi (talk) 14:10, 20 April 2013 (UTC)
 * That is a DirSync Pro developers agreement. Thank you, :-). I wait for some more agreements. UnTrueOrUnSimplified (talk) 14:33, 20 April 2013 (UTC)


 * I agree on merging the two columns and adding notes should a product handle moves differently than renames. Only one product has a different answer for "rename" vs. "move." I disagree on deleting both columns. Your assumption that all programs detect renamed and moved files is not true. For example, I know robocopy does not. If you rename a file and run robocopy again it will copy the file data to the destination under the new name. A product that detects renames/moves would see that the file is already on the destination but under a new name and/or location and just rename and/or move it on the destination. It would not copy the file data over again. Some products keep a "recycle bin" of deleted files and allow a person to delete a file and later to put it back. The product will see that the file data is already in its "recycle bin" and restore it rather than copying the file data. I am puzzled by the remark for iFolder where it says "Yes (files, not folders)." If a product can detect/handle files getting moved then by definition it can handle folders being renamed or moved. The product would simply see this as a bunch of files getting moved. --Marc Kupper&#124;talk 01:59, 7 May 2013 (UTC)


 * Given that the two columns are identical along with the preceding conversation, why don't we just merge them? If someone else doesn't beat me to it, I'll do it. Syzygy84 (talk) 21:10, 21 May 2014 (UTC)

Remark on "detect rename/move"
None of the open source software listed in the table (with the exception of Unison) has a true file rename/move detectability feature. If a file A was renamed to B, a software with this feature should detect that file B on the first server/folder is file A on the second server/folder. However, most of the listed software can only detect that file A is missing on the first server/folder, and file B is missing on the second server/folder. Therefore, stating that those software has rename/move detectability feature is misleading. I'd advise using a third option – Partial – which mark that the software can only detect if a file is missing, but not if a file was moved. – Fuzzy – 12:14, 9 July 2013 (UTC)

I think the definition above the table is quite clear what Renames/Moves is supposed to be: mimic the operation on the other side. Handling them the same way as new/removed files (and therefore just copying it over) should be marked as "No", there is no need for "Partial". OK, maybe I missed something, so what would be the difference between a "No" and "Partial"? I tested the behavior of FreeFileSync, and can confirm that it does not detect moves and renames, as required by the definition here, it simply threats them as new files on one side, and to be deleted files on the other. However I am reluctant to change this, since I don't have any sources, just my own testing. Any advice what can be done here? - Waldemar7k (talk) 22:07, 20 December 2014 (UTC)

More important (to me) than the efficiency gained by renaming/moving instead of achieving the same end state by copying/deleting, is the ability to do a proper bi-directional sync. in other words, it knows the difference between these two scenarios: 1) File A is present on Directory 1 but not on Directory 2 because it was deleted from Directory 2 (therefore delete it from Directory 1) 2) File A is present on Directory 1 but not on Directory 2 because it was added to Directory 1 (therefore copy it to Directory 2) To achieve this, it must hold metadata between runs — Preceding unsigned comment added by 165.120.27.234 (talk) 19:13, 21 December 2014 (UTC)


 * to the last poster (IP): Actually the modification dates of files and folders would make it posible (thou risky) to decide between "adjust by deleting" and "adjust by copying" without a database. Anyway the need for a Database (and complying use of the program) seems an important point in the threads main subject "reproduce moves and renames in target" vs. "delete old path, copy to new place/name from other side".


 * to Waldemar7k. True that Freefilesync (very good program anyway, like old win filesync) doesn't do the Job thou it states to use a database to recognize moves and conflicts. Even with Unison I didn't get it running (some time ago, must have been a bug), only time I really could reorganize my files without copying all of it again for just some minor name edits was when I was using commercial SuperFlexible File synchronizer.


 * Actually I think that we should have kept the extra columnas right now it leads to very wrong expectations. But maybe it could be redesigned and we can find a way with little columns, maybe including the "conflict" column. Perhaps we can find some 3 or 4 types of capability concerning comparison and analysis.


 * Complicated issue but important because I think it's very helpful for developers to know where they stand in the crowd of other solutions. --Dan-yell (talk) 00:08, 8 July 2015 (UTC)

to the unsigned comment (IP): this might be another issue, and more important to some people, but the comment does not belong here, there should be a spearate section for discussing two-way snycronization. That beeing said, I agree that it is a featrue that should be in the table as well.

to Dan-yell (and actually to myself too): I tested FreeFileSync again, more thoroughly now, and found out that it actually does the move/rename as said, but it's a little complicated, that's probably why I missed it the first time. First, it has to be turned on specifically, and the checkbox (called "Detect moved files") is only present in the "Synchronisation" tab of the options, not in the "Compare" tab. Second, this works only after at least one pass has been done before, with this option turned on. What happens the first time is, some kind of database is written, which seems to be used for subsequent passes to determine what will be moved (instead of copied again). There are a few problems with this system overall, for example, if more than one candidate is present to use for renaming (can be simulated by copying a file a few times), it seems to pick a random one, and doesn't seem to be aware of some conflict there. To be fair, it should be said that there is a tooltip that mentions both the database and that it's not available on the first run, it's still easy to miss if you do only test runs (comparisons that is) before actually synchronizing once and going on testing.

In general, I have to say I am too biased now to edit the main page anyhow, since I started to develop a program that handles such problems (will soon make it public, I think). -- Waldemar7k (talk) 18:42, 7 December 2015 (UTC)

Commercial vs Proprietary
The Commercial software section has the description of "This is a comparison of proprietary software released as commercial." Software can be commercial (sold) while still being open-source, and there's not a section for commercial open source software. I bring this up because "ownCloud" is in the commercial section, but it's open-source. It's actually using a dual-license monetization scheme, so I have no idea what section it should be under. Both? — Preceding unsigned comment added by 67.11.62.236 (talk) 14:00, 16 June 2013 (UTC)

FreeFileSync
FreeFileSync needs to be added. — Preceding unsigned comment added by 98.155.42.93 (talk) 16:54, 28 July 2013 (UTC)
 * It was eventually added. -79.177.122.137 (talk) 22:55, 2 May 2014 (UTC)

FreeFileSync seems to contain an adware virustotal.com scan — Preceding unsigned comment added by 176.147.136.169 (talk) 11:21, 21 October 2017 (UTC)

FTPbox
FTPbox -- open-source -- should be added, still in beta though. -> link: http://ftpbox.org/ — Preceding unsigned comment added by Tielemans.jorim (talk • contribs) 07:41, 8 June 2014 (UTC)

Critical Missing Table Attribute: Ability to Detect Changes in Padded Files
I apologize deeply if this critical attribute is already prominently addressed in the tables and I missed it, but if not...

I think it's very important to address the capabilities/shortcomings of applications that cannot detect differences in padded files and take appropriate or reasonable user-defined actions.

By "padded', I mean files that deliberately have extra space already included in them to accommodate rapid updates without requiring assessment of hard drive resources. The technique of padding is ubiquitous, if not standard, in application configuration files and audio files like MP3 and FLAC files, and has been so for well over a decade if not longer.

While the uninitiated may think this is a nit, a technicality or something that only the obsessive worry about, then I can reliably inform you that the real situation is completely the opposite for the typical user. For example, what if you edit the name of an artist in your music file and then save it to your local hard drive. Upon backing up or syncing your music files, would you expect that your change would be recognized and the appropriate update occur? I would hope so. But no...for many of the apps in this wiki article. Not only would the change not be detected, but many of these apps will see a file mod date change and no file size change, automatically assume the file dates where wrong, then overwrite the file mod date of the backup location to mask the detection of a difference. While in decades prior to 1990 this may have been a reasonable default action because of operating system and network issues in working with backups and external hard drives, in reality such issues were solved before even the MP3 codec was invented.

For the apps that can;t handle this well, they typically only have one way to detect real differences in added files: run the app by recalculating all checksum data on all files. So what would typically be a 1 minute effort becomes a multi-day effort. Or force the user to manually select files they somehow "knew" changed but weren't detected as changed to force an update. Neither are reasonable day-to-day. Put another way, one can always manually wipe out an entire backup or sync location and start from scratch each time that will probably take less time than a full checksum backup for sync effort, but no one needs an app for the method of wiping and redoing from scratch every time.

To be clear, this affects any user trying to not only deal with song files, but anyone trying to backup application configuration files - padding is an effective way to speed up small file changes without engaging in OS resource allocation and file moving efforts that can really slow down small configuration changes. I suspect many mainstream PC users use applications with configurations that they don't want to lose and therefore backup or sync, but if any are padded as is typical, any backup or sync app not designed to deal with padded files will at some point likely have a major issue.

I'm just a typical PC user that got burned by this when I discovered key files on my backup were very old and missing many metadata changes despite having a file modification date as recent as my actual latest master file that had all the latest metadata changes. I have absolutely no commercial interest in any application in this space. For that reason, I won't even list the apps that have this critical shortcoming or deal with it effectively and efficiently that I know of.

I heartily encourage someone with industry knowledge and/or wiki editing cred to enhance the tables to reflect what apps deal with this well, and which ones don't. Or at least note in the article that this critical feature is missing in the tables for now.

Thanks for considering this.
 * There are already too many columns in the tables, although this could be solved by removing the extremely sparsely populated and questionably useful ones (application and protocol layer; wow!  4 programs use TCP?  Jerkoff motion initiated!), and probably the "programming language" column, since this only matters if the language itself is required to use the program.  I have no idea what software you were using, but most software assumes the modification times are correct and overwrites the older (or newer even) file unless you tell it not to.  Robocopy works like that and has been included with Windows forever.   There's also the archive bit.
 * You never mentioned what OS you're using, so I'm just assuming "the most common one".
 * Any backup software that ignores modification dates and relies on file size should be considered fundamentally broken, anyway. Aside from your use case, which isn't how that normally works in the case of MP3, anyway (I think it's just part of the FLAC format so doesn't really count as padding);  try adding album metadata to a bunch of MP3s that didn't have any or use different formats of tag data and it'll immediately start increasing file sizes, I'm pretty sure the tags are just fixed-size data structures so once any of the information is there so is all of the space for it...  you can't have arbitrarily long artist / album names in normal MP3 metadata; you can have sparse files which are pre-allocated to some set size (possibly gigantic) but only occupy disk space for non-zero portions of the data, and files on NTFS can have alternate streams which can be used for practically anything and don't occupy space as far as the information shown to the user and badly written software are concerned.  It's not hard to make a 1kb text file that has a bluray ISO attached as Movie.txt:movie:$DATA or whatever even if doing so is kinda pointless.  Anything going on size alone would fail to backup either of those, and those things contain important data.
 * Assuming that programmers inserted padding in any file format because they had a single clue how the OS needed to handle data appended to a file (or OS resources when talking about extra data tacked onto music files) is vastly overestimating programmers. It's usually better to assume that their godawfully written spaghetti C code was initially intended to support strings of any size for the fields of the tag but they had no idea how or when to allocate and free memory correctly and kept OOM or segfaulting with their intial version.  If the person who came up with the tag type knew what they were doing, they probably thought about it for 5 minutes and realized that the 30 open source media players that would need to read this information from their godawfully written spaghetti C players would probably end up creating all kinds of vulnerabilities if they were forced to deal with variable length strings.
 * Without looking I'd assume that every single piece of software in this table does what you're asking, and you were just very slightly burned by using something completely horrible that you failed to mention the name of.  A Shortfall Of Gravitas (talk) 10:16, 31 October 2022 (UTC)

Missing some vital info -- Method and Scenario
When choosing a file sync software, some of the most important things to know are what method the software uses for achieving file sync (hashing vs attributes) and what scenarios its designed for (network vs local) or (always-on vs on-demand) or perhaps (peer-to-peer vs client-server). Very little of this info is detailed in the "Other" column, though often is missing altogether. There needs to be one or more columns that detail this info.

For example:


 * Syncthing uses file hashing, which is very CPU and disk intensive, but allows the app to compare file contents, and this is useful when you want to minimize network traffic, which is great for syncing network connected devices. However, this app is not designed for syncing local disks and folders on a single PC. It can't copy locked files (no Shadow Copy ability) and its only really designed for authenticated and wire-encrypted folder syncing which is too cumbersome to configure for local file syncing. This app is designed for always-on live syncing.


 * FreeFileSync compares file attributes (dates and sizes), which is very fast to do, and is perfect for syncing local disks and folders on a single PC. It can copy locked files, and offers batch configurations for common tasks like syncing multiple folders, filtering by various file attributes, and providing detailed sync stats. This app is not designed for syncing between multiple devices, and doesn't have any network awareness or wire-encryption or authentication features. This app is designed for on-demand or event-driven syncing. — Preceding unsigned comment added by 71.126.181.235 (talk) 21:35, 19 February 2016 (UTC)
 * I don't think Wikipedia is really meant to be a "help you pick which software you need to be running on your computer for each task" guide.  I'm surprised the 3(!?) lists already have the ridiculous number of columns that they do, since everything in them is required to have its own Wikipedia page (apparently that's notability) where presumably all of that information is explained in more details and it should really be up to the person reading it to research nitpicking details like "programming language" if they couldn't just use google to find the software they wanted in the first place...   As it is they're impossible to parse.  Should we add a "design philosophy" column too?  What about "source has doxygen comments"?   How about "ethnic, sexual orientation, and religious demographic makeup of programming team"? A Shortfall Of Gravitas (talk) 10:56, 31 October 2022 (UTC)

External links modified
Hello fellow Wikipedians,

I have just modified one external link on Comparison of file synchronization software. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:
 * Added archive https://web.archive.org/web/20110809233713/https://one.ubuntu.com/downloads/ to https://one.ubuntu.com/downloads/

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

Cheers.— InternetArchiveBot  (Report bug) 17:02, 11 August 2017 (UTC)

FreeFileSync
See Talk:FreeFileSync --Guy Macon (talk) 05:26, 5 February 2018 (UTC)

Cloud support: now it's very common
The table is not clear, you should add a column "sync could? yes/no".

Because, for instance, FreeFileSync doesn't sync clouds like Google Drive natively (you can use a workarond by installing a software that mount Google Drive a virtual drive in your computer), and some other softares in the table allow that.

It's criteria very important to chose a sync software.

Transport layer column
Add a table column for transport layers supported such as QUIC or DCCP.

Supported operating systems re-check...
I noticed that rsync is listed as having Windows support, but it doesn't. Being able to run under Cygwin (it's actually just distributed as part of Cygwin) still means installing the horrible hack that is Cygwin which isn't something anyone sane should be encouraging, and it definitely isn't Windows support. Might as well state that anything that runs on Linux has Windows support since you can just install WSL and bash on windows if you want to run linux ELFs from the command line...  Unfortunately I don't know how to phrase this. The list is also heavily biased towards GPL'd stuff, there's a bunch of MIT / BSD licensed file sync stuff floating around online, some of which has been around for decades, but as they don't tend to be pushing it like a religion there aren't wikipedia pages for each piece of software (and I don't have time to create them then fervently argue that they're notable enough for a page until everybody else gets sick of it and gives up) which prevents them from being added. A Shortfall Of Gravitas (talk) 10:50, 31 October 2022 (UTC)