The distinction between one-to-one binding and binding pools is the part I found most practical — strict binding makes sense for fixed meters and routers, but for cameras and trackers that get field-swapped, a pool with the spare IMEI whitelisted before the visit is clearly the workable option. The point that most lock incidents are self-inflicted also rings true: a SIM that still registers but silently drops data is a confusing failure mode if the technician does not know a lock is in place. I was also struck by the limits you flag honestly — a stolen whole device keeps a valid pair, so the lock is really about stopping SIM reuse rather than device theft. One question: in practice, how do teams handle the roaming case, where the check may behave differently on a visited network — do you treat roaming as an exception path in the policy, or avoid relying on the lock there at all?