Re: [Sks-devel] IPv6 peering; keydumps annoyingly large

[Matthew Palmer]

On Wed, Jun 01, 2011 at 06:14:34PM +0100, Xian Stannard wrote:
[redundancy through partitioning]
> Servers could carry multiple subsets to make sure that no particular
> subset lacked in redundancy?

But who coordinates which servers carry which subset(s)?  Without that, I
can pretty much guarantee that over time there would be some subsets that
would end up underserviced and at risk of being lost entirely.  This is
*particularly* true of something like the SKS keyserver network, where
individual servers are an unknown quantity regarding longevity and reliability
(and which would become even less reliable if more people on
resource-constrained systems were to become keyserver operators).

> I'm guessing that one of the design aims is that the network of servers
> needs to be redundant enough so that it is very hard to kill enough of
> them to start losing access to keys.

Indeed.  An analysis of how redundant the network needs to be would be much
appreciated before we go any further on the design of such a system.

- Matt

-- 
There's a term for those who fantasize that the world works in precisely the
way that produces maximum convenience for them, despite years of evidence to
the contrary.  The term is "Morons".
                -- Greg Andrews, in the Monastery