On Thu, Feb 25, 2010 at 11:00 AM, neil.young <[email protected]> wrote:
> Is there any paper describing the procedures/principles of node_id
> generation anywhere?

No, it's just common knowledge that the chance that two random 128-bit numbers
will be the same is negligible.


> One should add the term "statistically" to the globally
> unique claim, in order to prevent confusions.

I don't actually agree with this. Any method of generating unique values
has some nonzero chance of producing duplicates. For instance, if you're
using a counter, your memory might be reset to zero by ionizing
radiation.

-Ekr
_______________________________________________
P2PSIP mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/p2psip

Reply via email to