I know sk109340 addresses this exact issue, but I am still a little confused about the topic after reading it. The use of Empty Group Object in encryption domain is an confusing matter to me. This is the scenario I'm facing.
I have a pair of Cloudguard HA Cluster virtual gateways, and they have a couple Route Based site-to-site VPNs set up, with Virtual Tunnel Interfaces (VTIs.)
An Empty Group Object is created in the policy and that same Empty Group Object is being used as the Encryption Domain on both the Interoperable Device remote gateways, and the on the local Gateways. Everything is working with this configuration.
Now I have to set up a new VPN on this Cluster, to a different peer, that will be a Domain Based VPN (from what I understand, this basically just means a VPN without Virtual Tunnel Interfaces, and without the "Internal Clear" stuff in the policy rule.)
I believe what has to be done is removing the Empty Group Object from Encryption Domain of the local gateway, putting a new Group Object in as the Encryption Domain with the endpoints the new Domain Based VPN peer will be reaching to, and leaving the Encryption Domain of the Existing Route Based VPN Interoperable Objects as the Empty Group Objects.
My biggest concern is provisioning this new VPN without causing the existing ones to stop passing traffic. I don't like the possible scenario where after installing policy, the existing VPNs may start getting "according to policy we shouldn't decrypt the traffic" since I changed that local Encryption Domain to not be an Empty Group Object any more.
My lack of understanding about using an Empty Group in Encryption Domain is causing some doubt. Is there a good resource that explains why Empty Group is used in Check Point and what the difference is between using an Empty Group or a non empty one for Encryption Domains?