Before I make the GPS post below, I have a question for John: I heard on the news a few days ago that GPS was showing 20 to 30 m of motion of Sumatra and/or the Islands to the north. The way I understand a megathrust earthquake like this one, 1960 Chile, or 1964 Alaska...the subducting oceanic lithosphere (slab) would continuously sink into the subduction zone..flexing the hanging-wall (above the fault)......so that during the earthquake, the hanging-wall (in this case northern Sumatra and the Nicobar Island) would move suddenly to the southwest, while the Indian plate would not move very much.
An earth scientist I know thinks it would be otherwise...that the subducting plate would move the most during a quake.

Once you answer this, then the question is: is it the same when an oceanic plate is subducting beneath another oceanic plate? (I would think yes, the hanging-wall above the fault would move the most).

This is all relative to some fixed coordinate like the earth's rotation axis.
Chris