The way I think about it, there are three configurations for basic one-way file synchronization. I have made up names for the three.
- Push file synchronization. The files to be backed up are on computer A. The disks that the files will be backed up to reside in or on computer B. the software doing the backup runs on computer A.
- Pull file synchronization. The files to be backed up are on computer A. The disks that the files will be backed up to reside in or on computer B. the software doing the backup runs on computer B.
- Third-party file synchronization. The files to be backed up are on computer A. The disks that the files will be backed up to reside in or on computer B. the software doing the backup runs on computer C.
Sounds complicated, doesn’t it? It’s actually worse than that. Since we’re talking about more than two copies of each file, we can have hybrid approaches that use, say, push file synchronization from a workstation to a NAS box, third-party file synchronization with a server transferring data from one NAS box to another, and pull file synchronization with a server pulling data from a NAS box and transferring it to local storage which is then rotated to the safe deposit box. That’s not just a pedagogical way to look at things; in fact, it’s the scheme I use myself.
Here are some places where each type of synchronization might make sense.
- Push from workstation to server. You spend most of your time at your workstation. If the synch software is running there, you are likely to see if something’s wrong. You can push to several servers for redundancy.
- Push from workstation to NAS box. Advantages as above. You can push to several NAS boxes for redundancy.
- Pull from workstation to server. If you have many workstations, you only need one synch program to pull from all of them, and thus only one workstation backup program to keep track of. If you are providing backup for people who are technically unsophisticated, you don’t have to install any software on their workstations, and they don’t need to keep track of anything. You can have several servers do this for redundancy.
- Pull from workstation to NAS box. This would be a good solution in the case of many workstations if you could get good synching software for the NAS box, but I’ve not seen any.
- Third-party from workstation to NAS box with software running on a server. All of the advantages of pulling from workstation to server, but the data ends up on a NAS box. You can do this with several NAS boxes all from the same server, for redundant storage, or from different servers, for redundant storage and synching.
- Push from server to NAS box. This is a way to create redundant copies without taking up workstation resources for the second and third copies. When combined with pulling from workstation to server, you could do everything with only one file synch program installation.
- Pull from NAS box to server. Good for creating backups on a server that will be taken offsite. Backups can then be in a non-proprietary format easily read by the server or by workstations.