3 ms·

I will politely disagree. RSA-2048 implemented properly, on its own, has a significant security margin today and it's unlikely that HW advances over the next 25

by rphlx 11y ago

I will politely disagree. RSA-2048 implemented properly, on its own, has a significant security margin today and it's unlikely that HW advances over the next 25 years will erode that, even at a superbly-funded first-world intelligence agency. I am not saying that they won't be able to pwn you; just that they won't do so by directly cracking 2k+ RSA.

Spooky23 11y ago
The point is that crypto is only evaluated as "secure" at a given point in time. Even if RSA-2048 is secure in 25 years, it will almost certainly not be in the same place in it's lifecycle. Even today, NSA guidance recommends RSA-3072, or RSA-2048 with an accelerated migration to Quantum-resistant crypto in the future. (https://www.nsa.gov/ia/programs/suiteb_cryptography https://www.nsa.gov/ia/programs/suiteb_cryptography) My fear is that there won't be an "AES 2", and the commercial/individual world will return to a place where we don't have access to quality, trustworthy crypto for communications and commerce. Without security, there is no trust, and without trust many of the advances in productivity and and collaboration that we've gained in the last 20 years will be substantially weakened.