RE: Smart card proposal

["Lyal Collins" <lyal.collins () key2it com au>]

IMHO, the security issues of 'web sites communicating with tokens' are
poorly addressed.

If the site is hacker controlled (the main page site, advertising, image
sources etc), then anything can happen, good or bad.  While we want to good
to occur, the bad is just as important to control
Tokens effectively have a 'power of attorney' over one's authenticated
services.
Being realtively, dumb, they follow a few simple rules to generate output
that commits the human to the consequences of the input.

Human-mediated input/output, and indeed control over the 'power of attorney'
processing is severely lacking in the products available today.  No
solutions appear to provide strong, ongoing control by the user of how, when
and why the device is used to authenticate/encrypt transaction related
content.

A refocussing on products that address business and user needs is necessary.
I fear that most infrastructure and products in this space is too insecure
or dumb to actually do any of the necessary  human mediated secure
communication and authentication

Lyal
 

> -----Original Message-----
> From: Koh Gim Leng [mailto:kohgimleng () gmail com] 
> Sent: Friday, 28 January 2005 2:10 PM
> To: Richard M. Smith
> Cc: webappsec () securityfocus com; maburns () safenet-inc com
> Subject: Re: Smart card proposal
>
>
> On Tue, 25 Jan 2005 16:45:52 -0500, Richard M. Smith
> <rms () computerbytesman com> wrote:
> > Hi,
> >
> > What seems more user-friendly to me is that a single USB 
> key device needs to
> > be sharable between 5 credit cards, rather than having 5 
> separate devices.
> > What I still do not understand is how a Web site 
> communicates with the USB
> > key device.
>
> Web site communicates to a web browser which in term communicates to a
> set of standard APIs known as PKCS#11. Every smart card token or USB
> token vendor who wants to sell their token will provide you with their
> PKCS#11. Being a standard, your browser will know which API to call
> within the PKCS#11.
>
> If the token is in the form smart card, then the PKCS#11 provided by
> the smart card vendor will include means to call the smart card
> drivers provided as well. If the token is in the form of USB, then the
> PKCS#11 provided by the USB vendor will inculdes means to call the USB
> drivers as well. Bottom line - whatever happens belows PKCS#1l should
> be transparent to the web browser (or any PKI enabled application).
>
> Regards
> Sam Koh