[Bitcoin-development] Draft BIP for Bloom filtering

[Mike Hearn]

> Because I can potentially waste bandwidth of all nodes forever (well as long
> as users are still scanning blocks with my transactions in them) with O(1) work.

And this gets you what?

Users who have active wallets will have their bandwidth wasted for as
long as you keep up the attack. Once you stop active wallets won't be
rescanning that part of the chain and new users won't be scanning it
either, as they skip blocks before their earliest key time using
getheaders. So basically you can waste the bandwidth of active users
for a while, by spamming transactions. This is not a new attack.

Anyway, it's trivial to DoS the entire Bitcoin network today. It
hasn't ever happened. Maybe one day it will, but the only rationale
people can come up with for such an attack beyond random griefing is
governments, and complexity attacks are really not their style. Much
easier to just pass a law.

I'm not saying DoS should be ignored, but I do feel there are limits
to how far down that rabbithole it's worth going.