Is there any paper describing the procedures/principles of node_id
generation anywhere? One should add the term "statistically" to the
globally unique claim, in order to prevent confusions.
Regards
Regards
Eric Rescorla schrieb:
On Thu, Feb 25, 2010 at 8:46 AM, neil.young <[email protected]> wrote:
Hi Eric,
<snip>
In most cases, there's a centralized enrollment authority which doesn't
issue duplicates.
</snip>
Yes, I believe your claim for self-signed certs. But a centralized
enrollment server in an overlay which issues certs and guarantees no
duplicates can't always guarantee global uniqueness. It is likely, but not
literally guaranteed. The "globally" claim is what I doubt :)
This actually is not correct.
If the authority pseudorandomly generates them (which it should) then the
nodeids will in fact be statistically globally unique.
-Ekr
_______________________________________________
P2PSIP mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/p2psip
