Currently, power management features can affect multicast link speeds. If a client is in a low power state, even if it is not part of a ghostcast session, it can cause the entire imaging job to run at very low speeds(have seen 5MBs).
The issue seems to be when ghostcast is polling for clients, these low power state clients report their low linkspeed, eventhough after waking up, they are capable of much faster speeds, Ghost does not recheck to if that was the case.
The enhancement request is for ghost to not only repoll for linkspeed, but to be able to set a minimum threshold so that if an imaging session cannot be performed at 'X' MB/s, the job doesn't start at all.
With images reaching 30GB today, if a job is set to run from the Console on Friday night for computers to be ready for Monday, the job will not be ready and none of the clients will be usable after that 48-60 hour time period.