CDR: Re: Public Key Infrastructure: An Artifact...

[Greg Broiles]

On Mon, Nov 20, 2000 at 02:18:42PM -0600, Jim Choate wrote:
> Real-To:  Jim Choate <ravage at ssz.com>
> 
> 
> On Mon, 20 Nov 2000, R. A. Hettinga wrote:
> 
> > At 12:10 PM -0500 on 11/20/00, Arnold G. Reinhold wrote:
> >  
> > > If CAs
> > > included a financial guarantee of whatever it is they are asserting
> > > when they issue a certificate, then all these problems would go away.
> > 
> > Right.
> 
> Bonding would not fix this problem. It only moves the question of identity
> and responsibility to the bonding agency. You've still solved nothing.

It's not a bond; and it doesn't solve the problem directly, but moves
responsibility for solving the problem out of the end users' domain and
into the CA's (or guarantor's) domain, where their greater resources and
experience (and liability) will help them solve the problem in the most
efficient and economic fashion.

It's like putting prices on corporate or government bonds - you can look
at the price of the bond to get an idea of the confidence people have in
the likelihood that the underlying obligation will be repaid. 

Certificates which are priced on a risk-sensitive basis - or whose 
face value (or guarantee value, or whatever) is risk-sensitive allow
people (and their computers) to immediately see both their own risk
exposure in concrete terms, and to have an idea of what the market
(including sophisticated participants) thinks about the risk.

--
Greg Broiles gbroiles at netbox.com
PO Box 897
Oakland CA 94604