Transaction prioritization[1] and transaction fees already rate limit stuff like that, and patches to the clients used by major mining pools may be able to fix it permanently.
1. priority = sum(input_value_in_base_units * input_age)/size_in_bytes
1. priority = sum(input_value_in_base_units * input_age)/size_in_bytes
In general, let's say a govt agency get assigned an $18.5m budget to break the Bitcoin currency as much possible, what could their plan of attack be, i.e. spend that money in the most efficient way possible? Create a large ring of wallets and send tiny payment around the ring? Create less wallets but send large payments between them?