[Cryptography] quantum computers & crypto

Jerry Leichter leichter at lrw.com
Sun Oct 31 13:07:59 EDT 2021


New Scientist (https://www.newscientist.com/article/2293341-spies-may-be-storing-data-to-decrypt-with-a-future-quantum-computer/) brings up another issue with quantum cryptography:  Some secrets may remain valuable for years or even decades.  Intelligence services have a long history of saving encrypted material that they can't break in the hopes that they will be able to break it eventually.  The Venona decrypts are probably the most famous example of this kind of thing.

Brought up in the article is the possible connection to repeated BGP misconfigurations that cause large amounts of traffic to transit Russia and China - and of course even in normal circumstances much of the world's traffic transits the US.  Are these countries "saving it all" for a post-quantum future?

Dealing with this kind of thing requires some careful analysis of the kinds of data being transmitted.  Most transactional data becomes irrelevant not long after it's sent.  Knowing how to break into Swift transactions, much less the vast majority of on-line Web transactions, 10 years after the fact has almost no real value.  For intelligence services ... it's always of some value to figure out who was active even 10 years ago, as connections continue into the present. Maybe you can find blackmail material.  Maybe you can grab passwords that have stayed unchanged over all this time.  But for the most part ... who cares.

There is certainly government/military information that remains valuable after a decade.  Is there much *commercial* information still valuable after that long?

Suppose we reduce the time period to 5 years, or 1 year.  How does the picture change?  Passwords certainly do survive for a year - but then again, we really shouldn't be sending passwords around anyway - we have plenty of protocols that do "proof of knowledge" without sending the actual secret.  How do those hold up in a post-quantum world?

Symmetric algorithms seem fine, even in a post-quantum world.  Much of the most valuable material is non-transactional and is encrypted at rest.  Even if you send it over a channel that can be broken in the future ... it doesn't matter.  But ... watch out for the common technique of attaching the symmetric key encrypted with multiple (asymmetric) public keys to allow access to multiple individuals.
                                                        -- Jerry



More information about the cryptography mailing list