How about PC vs PC sync other network?

[rgreat]

I mean not through shared folder, as it is very slow to scan, but between 2 FreeFileSync applications?

Each act as local folders scanner and then they compare results between each other.
That would be many times faster.

[xCSxXenon]

That wouldn't be any faster than shared folders. I am using shared folders at over 1 gigabyte per second and have no issue using FFS or native File Explorer.
If you are only referring to the comparison phase, it still would have to transfer the file info one way or another for the FFS sessions to compare to its own scan. Still not faster, probably even slower

[rgreat]

That wouldn't be any faster than shared folders. I am using shared folders at over 1 gigabyte per second and have no issue using FFS or native File Explorer.

I'm glad that you have such a nice connection, but not everyone have it.
For me remote scans are like 100+ times slower (mostly due to latency).
I need to sync folders from server to my PC.
I use RDP over VPN/internet (with limited network access) and share "Local Resources" to have access to my PC disks from server.
Have to scan large directory (7000+ subfolders and 75000 files), out of which i usually only need to transfer like 100 files and up to ~100 mb of data.
Local scan (on server) is quick, transfer to my PC is quick, but scanning my PC files remotely from server take a hour or so.

If you are only referring to the comparison phase, it still would have to transfer the file info one way or another for the FFS sessions to compare to its own scan. Still not faster, probably even slower

Whole idea is to exclude fire transfer (and frequently repeated operations) over the network.
Remote FFS would do it locally (on remote server) and only transfer results (databases), like file lists, timestamps and CRC's after it complete a full scan.
After that local FFS will compare remote database with local one and initiate file transfer.

Should be much faster. Like 100-1000 times faster.

[Plerry]

Your proposal would require two instances of FFS, each running on respectively the left- and right location (for a sync job that has just a single left-right pair).
If you are syncing between two computers you have full control over, that would theoretically be an option.
However, most of the time people run syncs between a computer (over which they have control and) on which they can run FFS, and an external resource, like a USB disk, a network location (e.g. a NAS) or a server over which they have no control beyond file-access (e.g. a domain server or cloud service). Running an FFS instance on any such external resource is then not possible or not allowed.

For that reason, the architecture of FFS is such, that you (need to) run just a single instance of FFS on any suitable machine that has file access to you left- and right location(s). Your left- and/or right-location(s) may, but does not need to, reside on the machine running FFS. This allows you even to sync between two external resources.

[rgreat]

Your proposal would require two instances of FFS, each running on respectively the left- and right location (for a sync job that has just a single left-right pair).
If you are syncing between two computers you have full control over, that would theoretically be an option.
However, most of the time people run syncs between a computer (over which they have control and) on which they can run FFS, and an external resource, like a USB disk, a network location (e.g. a NAS) or a server over which they have no control beyond file-access (e.g. a domain server or cloud service). Running an FFS instance on any such external resource is then not possible or not allowed.

For that reason, the architecture of FFS is such, that you (need to) run just a single instance of FFS on any suitable machine that has file access to you left- and right location(s). Your left- and/or right-location(s) may, but does not need to, reside on the machine running FFS. This allows you even to sync between two external resources.

Still, if FFS would have a new Master-Slave option added - it would be nice in some cases.

[hourglass]

Hello,

I would agree with rgreat.

There are many reasons for using freefilesync I'm sure and you know better than me the most frequent usage.

However, to sync my desktop and laptop twice a day when arriving/leaving the office (up to 300,000 files across 26,000 folders), running two instances (server/client type) would likely help to update the changes. I assume at least.
That scenario may be more frequent than before with home working some days of the week that is quite widespread now (at least in France).

So I can't use FFS for that although I would like to as I find it great when I use it to keep in sync results from simulations on linux servers that need to be processed elsewhere.

Thanks
Philippe