Hi I am investigating a use case with Syncthing where I am using multiple networks groups were each group is sharing a single folder. With each group in these networks there is 1 and the same client syncing between these networks. This client is not altering any data. See picture for more clarity.
Some metrics.
Per group: Up to 50 clients per group and up to 5 Gb per folder (group).
How many groups can be supported in this way for the client in the middle?
Is this a feasible use case for Syncthing?
How can you make sure the client is coping with all the changes and data, maybe special hardware?
I’m using somewhat scaled down (node-wise) version of the mentioned “groups” setup in one of the clusters. The setup consisted of <10 groups, near 10 users in each, with few tens of folders per group, and about 3TB of total data in sync. And there was one central node connected to all groups.
As far as my experience go, syncthing copes with it just fine. The central node should have some extra RAM, fast storage for DB, and a decent processor.
The only thing to watch for, is conflicting changes, if you have a big number of users, there is a chance that 2 users might make modifications in parallel (before each other’s changes propagate fully to all the nodes). Staying connected constantly helps to reduce conflict probability.
