A mighty fortress is our PKI

[Anne & Lynn Wheeler]

On 07/27/2010 12:09 PM, Pat Farrell wrote:
> In that same time, I was at CyberCash, we invented what "is now
> sometimes called "electronic commerce". " and that and $5 will get
> you a cup of coffee. We predated SSL by a few years. Used RSA768 to
> protect DES sessions, etc. Usual stuff.

somewhat as result of doing the SSL payment stuff ... in the mid-90s got invited to be part of the x9a10 financial standard working group ... which had been given the requirement to preserve the integrity of the financial infrastructure for all retail payments. the result was x9.59 retail payment financial standard ... which was specific in such a way that it would work with any secure authentication (including allowing both certificate & certificate-less mode). The business process was slightly tweaked so it was no longer necessary to hide the information in a payment transaction to preserve the financial infrastructure integrity. This didn't eliminate skimming, evesdropping, data breaches ... but it eliminated the ability for the attackers to use the information to perform fraudulent transactions (and effectively also eliminates the major use of SSL in the world ... hiding the information in financial transaction).

About the same time the x9a10 standards work was going on ... there were a couple other payment transaction specification work occurring ... which were mandating certificate operation ... somewhat trying to side-step the 100 times payload bloat. they would strip the certificate at internet gateway ... and forward the transaction thru the standard payment network with flag turned on
(they could somewhat wave their hands that 100 times payload bloat on the internet was immaterial ... but not so in the real payment network) that certificate processing had occurred (compared to light-weight, super secure, x9.59 ... which operated end-to-end). There were later some presentations at ISO standards meetings that transactions were showing up with the "certificate" flag on ... but they could prove no certificate had been involved (i.e. there was financial interchange fee benefit motivating turning on the flag).

shortly after they had published their (certificate-based) payment specification (but well before any operational code), I did a public-key op profile for their specification. I then got a friend that had a optimized BSAFE library (ran four times faster) to benchmark the profile on lots of different platforms ... and then reported the results to the groups publishing the profile. The response was my numbers were 100 times too slow (if they had actually run any numbers, their comment should have been it was four times too fast). Some six months later when they did have pilot code ... my profile numbers were within a couple percent of actual (i.e. the BSAFE library changes had been incorporated into standard distribution).

-- 
virtualization experience starting Jan1968, online at home since Mar1970

---------------------------------------------------------------------
The Cryptography Mailing List
Unsubscribe by sending "unsubscribe cryptography" to majordomo at metzdowd.com