https://wiki.deltaworlds.com/index.php?action=history&feed=atom&title=Aw_random 2026-04-17T01:58:16Z Revision history for this page on the wiki MediaWiki 1.39.3 https://wiki.deltaworlds.com/index.php?title=Aw_random&diff=198&oldid=prev 2025-04-07T19:26:01Z

<p>Created page with &quot;{{SDK added in|version=2.1|sdk=13|world=|universe=|browser=}} __NOTOC__ int aw_random (void) ==Description== Returns a random number. ==Notes== The psuedo-random number generator (PRNG) employed by this method is [http://www.math.sci.hiroshima-u.ac.jp/~m-mat/MT/emt.html Mersenne Twister] (MT) by Matsumoto and Nishimura. It sets new standards for the period, quality and speed of random number generators. The incredible period is 219937 - 1, a number with about 6000 d...&quot;</p> <p><b>New page</b></p><div>{{SDK added in|version=2.1|sdk=13|world=|universe=|browser=}}<br /> <br /> __NOTOC__<br /> <br /> int aw_random (void)<br /> <br /> ==Description==<br /> Returns a random number.<br /> <br /> ==Notes==<br /> The psuedo-random number generator (PRNG) employed by this method is [http://www.math.sci.hiroshima-u.ac.jp/~m-mat/MT/emt.html Mersenne Twister] (MT) by Matsumoto and Nishimura. It sets new standards for the period, quality and speed of random number generators. The incredible period is 219937 - 1, a number with about 6000 digits. The 32-bit random numbers exhibit best possible equidistribution properties in dimensions up to 623 and it&#039;s very fast.<br /> <br /> ===Predictability===<br /> This PRNG is should &#039;&#039;&#039;not&#039;&#039;&#039; be used in cryptography as it has major vulnerabilities:<br /> <br /> *Observing a sufficiently long sequence of outputs from MT is enough to predict all future outputs.<br /> *It takes many iterations before the initial non-random state produces output that will pass a randomness test.<br /> <br /> Two steps can be taken to mitigate the problems:<br /> <br /> *Generate multiple 32-bit outputs (e.g. 8 would be sufficient) and digest them using a [http://en.wikipedia.org/wiki/Cryptographic_hash_function cryptographic hash function] such as RIPEMD or SHA-1.<br /> *Throw away the first x number of outputs from MT (e.g. call aw_random a million times when the SDK application is initialized).<br /> <br /> ===Seed search attack===<br /> The implementation used in the SDK is also vulnurable to a &quot;seed search attack&quot; due to only using a 32-bit number for the seed (i.e. it can only produce 2^32 different sequences of psedo-random numbers). This means that attackers would need to try at most 2^32 different seeds before finding the one being used.<br /> <br /> If a large number of initial outputs of MT have been thrown away then attackers would need to generate a longer sequence of numbers to test each seed. If 1,000,000 numbers were thrown away and 1000 have been passed to users then an attacker would need to generate roughly 1001000 * 2^32 numbers to be sure of finding the correct seed (in addition to seeding 2^32 times).<br /> <br /> ==Arguments==<br /> None<br /> <br /> ==Argument attributes==<br /> None<br /> <br /> ==Return values==<br /> Signed 32-bit number (i.e. a range of -2147483648 to +2147483647).<br /> <br /> ==Returned attributes==<br /> None<br /> <br /> ==Usage==<br /> <br /> Naive<br /> <br /> if ([[aw_random]] () == 547891)<br /> puts (&quot;You sure are lucky&quot;);<br /> <br /> Better<br /> <br /> void random_init (void)<br /> {<br /> int i;<br /> <br /> for (i = 0; i &lt; 1000000; i++)<br /> [[aw_random]] ();<br /> }<br /> <br /> int random_secure (void)<br /> {<br /> int n;<br /> int i;<br /> <br /> n = 0;<br /> <br /> /*<br /> combine multiple outputs into a single number using xor<br /> */<br /> for (i = 0; i &lt; 8; i++) <br /> n ^= [[aw_random]] ();<br /> <br /> return n;<br /> }<br /> <br /> random_init (); <br /> <br /> if (random_secure () == 547891)<br /> puts (&quot;You sure are lucky&quot;);<br /> <br /> Secure<br /> <br /> void random_init (void)<br /> {<br /> int i;<br /> <br /> for (i = 0; i &lt; 1000000; i++)<br /> [[aw_random]] ();<br /> }<br /> <br /> int random_secure (void)<br /> {<br /> int buf[8];<br /> int i;<br /> <br /> for (i = 0; i &lt; 8; i++) <br /> buf[i] = [[aw_random]] ();<br /> <br /> /*<br /> function that takes a buffer as input, digests it using RIPEMD-128<br /> and returns a 32-bit hash value from the output<br /> */<br /> <br /> return ripemd128_32 ((char*)&amp;buf, sizeof (buf));<br /> }<br /> <br /> random_init (); <br /> <br /> if (random_secure () == 547891)<br /> puts (&quot;You sure are lucky&quot;);<br /> <br /> ==See also==<br /> *[[aw_init]]<br /> <br /> [[Category: SDK Methods|R]]</div>

Xan