4 ms·
I like the argument that if you can break 2^100 complexity, you can probably break 2^200 because it will have meant you turned NP into P. But... I have a 64Mhz
by SlowRobotAhead 7y ago
I like the argument that if you can break 2^100 complexity, you can probably break 2^200 because it will have meant you turned NP into P.
But... I have a 64Mhz embedded device doing 10,000 ChaCha20-poly1305 encryption operations per second on 64 bytes of data and 64 of “add” in the AEAD MAC. Small packet, but those numbers are crazy.
Chacha is FAST considering it doesn’t need the same special hardware that AES does.
I guess I get the idea of not doing extra work and at a global scale it might add up. But I won’t move my rounds down from existing standard unless I needed to.
- Avamander 7y agoIMHO if we look at the global scale, we might as well optimize tons other software or operating systems before. Or we could force-enforce SPF and DMARC and kill a lot of e-mail spam and save power like that. Reducing cryptosystems' safety margins doesn't just seem optimal to me.
- SlowRobotAhead 7y agoTotally agree. I just figured out the other day that if I turn the POE off at our offices from 12am to 5am, it saves us $500 a year and a little durability on the switches. It's not processing, but I guarantee it's a larger difference for us than dropping rounds out of encryption.