I’m looking to build a private peer to peer data replication system (files). Essentially I need every host in the network to have every file that gets pushed in. Reading through the IPFS documentation, it sounds like a perfect fit for what I need. We have a lot of hosts but likely we’d split things up into several groups of isolated networks. Each network containing around 200 hosts (maybe less).
The data we need to move around is generally less than 100k a minute, but twice a day we have several gigabytes to move through the system.
The other major requirement we have is that we need essentially 100% up time. Each host must receive every file. If a host starts falling behind or fails to receive the data, it ends up having to be taken out of service and replaced with one that is bootstrapped with the latest data so that it can rejoin the network up to date. While we expect this from time to time due to bad network connections, failed hard disks, etc, it cannot happen simultaneously across all our hosts due to a software failure.
Can somebody elaborate on how production ready ipfs is, specifically private networks and guaranteed delivery. I’d like to better understand what our uptime is expected to be if we built out an ipfs private network?