Randomness - Part 1
I had a statistics class this weekend and our professor happened
to talk about "random number generators"(for clarity sake: random
number generators(RNG) is used to generate random numbers). He stated -
"all random number generators in this world are not truly random, they are
just pseudo-random". This is quite true since computers are basically
zero-intelligence elements and there would be a particular formula using which
it generates a random number. Hence however long the sequence is, after a
particular time period the sequence of numbers would start repeating.
Now this is a
real issue when you consider areas like cryptography and military. These fields
rely very heavily on the fact that the RNG's generate truly random numbers.
This vulnerability was recently exposed by a group of scientist who cracked the
formula/algorithm of a RNG(used for military purposes) with a minimal budget
and a short time span of just 57 hours.
This was a very
serious threat and had to be overcome quickly. So what does one do?
Continuing on
with my lecture this weekend, my professor then stated - "Only physical
events like tossing a coin, rolling a die etc are truly random in nature".
Voila!! So we have found out a way to generate truly random numbers. Such a
simple idea.
But integrating
physical real world phenomena like tossing of a coin into a computer is a
really tough task. It is equivalent of telling the computer - "Hey!!Here
take this coin and toss it a 100 times and let me know the set of
outcomes". So the computer experts had to find another way to do this.
One simple and most frequently used concept to generate random
numbers is by feeding the "system time" as an input to the RNG
function. Since the system time is dynamic, one can expect a set of random
numbers for a fairly long period of time. But again this is NOT truly random.
And so here is another way..What if we capture a real world
phenomenon like atmospheric noise, thunder, temperature sensitivity etc. This
is no rocket science and can be easily given as an input to the RNG. We all
know how random these atmospheric events are. Has a weather forecast made by
the met department ever be as they claim? Though they have a thousands of
terabytes of data on temperature, rain pattern for hundreds of years, they are
unable to forecast the weather accurately.
Hence we can conclude that if an atmospheric event can be fed as
an input to the RNG's, they would end up producing TRULY random numbers.
To be continued
in Randomness - Part 2
Vartmankal
(TT - 00:30)
Comments
Post a Comment