Question regarding Synth Full vs Active Full at remote sites

[$username]

So the situation we're looking at is we have a number of remote site, they all have a single ESXi host attached to our central vCenter, each of these hosts have local disk storage with one datastore for all VMs on it.

Right now we have one proxy running on these hosts with a WAN cache, our intention is to also spin up a local repo and carry one full with one day prior incremental and the do a backup copy job and carry one weekly full and the rest of the incrementals back to CO over the WAN.

So I am wondering what would be the best way to go about this, should we be using forward vs reverse incremental, synth vs full weeklies or is our idea totally wack and the wrong way of going about this?

Our overall goal is to be able to offer a quick restore of a VM if it died today and quick restore of recently damaged files, everything older than 48 hours can be handled out of CO.

[skrause]

If you are only keeping 2 restore points at the remote site, you probably will be best served by not running synthetic fulls, just doing forever forward incremental or reverse incremental. Forever forward is probably the better option since the snapshots are not open for nearly as long and then the transforms happen afterwards. Unless your change rate is really high, I would probably keep at least 5 restore points on local disk so that if someone deleted a file on Friday morning and doesn't tell you until Monday afternoon you could still have it be local