@devjit2018 Some good points.
For the record, overlapping doesn't always refer to conflicts. It can merely refer to inefficiency and redundancy... most of the anti-ransomware solutions are using bait files or monitoring file system behavior from a file system mini-filter. So most of the time, all you are doing is just wasting system resources, even if you're using many different solutions to cover the same thing which aren't causing real conflict. It would be like employing ten people to do a five people job where everyone is doing the same thing, even when they can do their jobs without conflict with one another.
Some of them might watch for file extension changes, dropping of ransomware notes, monitoring the use od vssadmin, etc. I've even seen cryptography libraries being hooked before in user-mode.
I would go for your last practice than the rollback idea. The rollback features are just another example of cat and mouse - will it work or will it fail?
If you have good backups of your data, there is low chance of failure. You can have several backups on the cloud - be sure to encrypt them appropriately - and offline. If for whatever one backup goes south, the others should be fine.
If your data is backed up and your prevention layers fail, you won't even have to consider the idea of paying a ransom or hoping that the data can be recovered by someone with a decrypter. Even worse if the ransomware was never designed to support recovery and thus by paying a ransom, you get nothing.
The rollback features have to store the original data somewhere. Even with compression, if you support rollback for a lot of data, it is going to consume disk space. It'a inevitable. I personally would just advise backups.