Xbox key defended by more than just length

Posted Jan 9, 2003 9:47 UTC (Thu) by beejaybee (guest, #1581)
Parent article: Xbox key defended by more than just length

Well, they aren't RSA keys, but NFSNET (http://www.nfsnet.org/) is routinely factoring numbers around 2^700 - at a rate of around one a month, without an immense number of volunteers and using hardware requiring only around 128 MBytes memory.

Highly sophisticated factoring methods like NFS scale much better than most people realise. Indeed factoring has been shown to be an "easy" problem, as a polynominal time algorithm has been published.

RSA 2048 may well be safe for a few years yet (at any rate, the commercial life of the Xbox) but don't simply assume that it is safe for the lifetime of the Universe.

As for the Neo project's aborted effort - their approach was indeed extremely unlikely to succeed - IMHO M$'s lawyers' advice _should_ have been to leave them in business, perhaps even provide a little covert support, in the hope of diverting effort away from projects which have a much better chance of success.

Xbox key defended by more than just length

Posted Jan 9, 2003 11:10 UTC (Thu) by beejaybee (guest, #1581) [Link]

Um.

I wrote "Indeed factoring has been shown to be an "easy" problem, as a polynominal time algorithm has been published." This is not quite true - there is indeed a polynominal time algorithm but it depends on quantum computing.

QC will, of course, totally destroy RSA, whatever the key length.