[Beowulf] OT: public random numbers?

David Mathog mathog at caltech.edu
Fri Aug 12 10:22:46 PDT 2011

Michael Di Domenico wrote:

> How many random numbers per day are you expecting?

One would be sufficient.

> If everyone checks at exactly 1pm, should they all see the same
> "random" number or should they each get their own "random" number?

They should all see the same number.

Example: a random number based on physical events which occurred on
8/10/11 would become available on or shortly after that day.  Starting
from the time it first becomes available, and going forward ideally
forever, everybody who wants to should be able to retrieve that same
random number.  

That is, nobody should be able to predict the number before hand, and
everybody should be able to verify it later.  So the number must be both
random and etched in stone.

> What kind of entropy are you expecting on "random"?

In practice relatively little is needed, 16 bits should be plenty.
(More wouldn't hurt, of course.)


David Mathog
mathog at caltech.edu
Manager, Sequence Analysis Facility, Biology Division, Caltech

More information about the Beowulf mailing list