Let’s suppose I build an IPSec tunnel between myself and my cloud provider. Assume my firewall’s IPSec configuration allows both IKEv1 and IKEv2, but IKEv2 is established with my peer by default.
Assume there is a bad actor in the middle of this tunnel. How feasible is it for an adversary to perform an IKEv2 to IKEv1 downgrade attack against this configuration if Phase 1 is already established using IKEv2? Also consider the possibility that I do maintenance which takes the tunnel down and then the attacker tries to force IKEv1 during tunnel re-establishment.
Good practice is to remove unwanted IKEv1 configurations; however, I would like to understand the feasibility of someone actively performing an IKEv2 to IKEv1 downgrade attack. Building IPSec tunnels in terraform on a well-known cloud provider puts you in this exact situation where the tunnel can be established with both IKEv1 or IKEv2 unless you explicitly disable IKEv1.