> Let us keep in mind that the NSA has computer systems > that are at > least a decade advanced beyond what we have in the commercial/civilian
How do you figure this? guess? > sector. This may seem unremarkable at first blush, but > consider that in > 1991, we were all running on machines that were in the > double-digit MHz > range at best. Today, the one GHz (~1,000 MHz) and better prevails. > Operating on the presumption that computing power advances by > a factor of > ten every five years[1], it becomes readily apparent the sort > of power the > NSA can potentially wield in brute-forcing any number of cryptographic > schemes. Hmm, so you are saying that the NSA moving forward at the same constant as everyone else is 10 years ahead? If they are 10 years ahead now, they would have to have been at least 100 years (or more) ahead 10 years ago. This is highly unlikely. Moore's Law refers to computing power overall. Everyone is moving at the same speed. Of course the fact that equipment moves down to the 'consumer' more rapidly now, than in the past; would mean that the 'consumer' is catching up to the NSA not the other way around. This is the norm, the lower end of the 'system' will always accelerate toward the top end. The fact is that as time goes by, the difference in years becomes compressed. They may be a year or two ahead, but 10 is not likely. In the early 1980s they would have been maybe 5 years ahead, but not enough to now be 10 years ahead. While the NSA does hire a large number of mathematicians and comp sci folks; they by no means have the corner on talent. It would difficult to believe that they had created a 'process' that would accelerate their code breaking efficiency to be 10 years ahead. Possible, but I feel not likely. > > 4. 2001 hasn't exactly been a banner year for Microsoft > IIS, Outlook, > et cetera. Open source is not holding up well either.
