encryption - Why is it impossible to implement an "infinite" one time pad algorithm? -
i've done research this, i'm still not sure why cannot implemented. provided share initial otp, possibly via usb or other physically secure method, surely can include next 1 in messages follow.
[edit: more specifically, if take pad of double length, splitting x , y. using x encrypt message, , using y twice encrypt next pad, insecure?]
you have pair each bit of message same size bit of otp. there's limited amount of otp.
if pair of otp bits bits next otp...
a b c d e ... q w e r t ... there's no room message. , if keep spending otp transferring otp, there never room message.
you can't compress otp, because strength of otp it's random - that's makes impossible codebreakers, because there's no pattern latch onto.
compression technology works finding patterns , replacing them shorter "that large repetitive block goes here , here , there" signals - , definition there no patterns in complete randomness, otps not compressible.
if can compress bit, it's not right describe otp anymore, it's weak - , massively wasteful of bandwidth. if can compress lot, throw random number generator away it's terrible.
quick test demonstration of concept on linux machine:
$ dd if=/dev/urandom of=/tmp/test count=10k -> 5mb file of randomness $ bzip2 /tmp/test -> 5.1mb file $ gzip /tmp/test -> 5.1mb file compressing pad makes bigger, adding bzip/gzip file format information , doing nothing else.
Comments
Post a Comment