Is side channel snooping really possible?
If side channel eaves dropping is possible, Laurie, there's no current in-place protections. It's only been done in the lab and some researchers say the initial group's findings are not repeatable. Nevertheless, I do not rule out the possibility it could occur. One VM listens for the physical activity that signals a virtual machine waking up to incoming traffic. If the neighbor can identify the target virtual machine -- a big if, once Amazon stopped numbering them in a predictable sequence -- then it listens for a keystroke pattern that might tell it the sequence of the first data in, the encryption key. Offhand, I would say this is nonsense, you can't accomplish all that. But stranger things have happened, One protection, not in place yet: send one or two initial false pulses of data, resembling a key, followed by the actual key. I think the idea is, by repeated listening to the sequence, a knowledgeable observer might piece together the key from the keystroke pattern. If so, that's a big exposure.