Re: Bag of Data Anti-pattern

On 01/04/2018 12:44 PM, Adrian Gropper wrote:
> Given Mike's comment with respect to biometrics, would we ever want
> a public biometric template in the DID document or is a template
> always proprietary and/or method-specific?

Let me attempt to clarify the biometrics use case because there seems to
be a fundamental misunderstanding here:

You should never put private data on a blockchain, EVER... that includes
raw biometric information. That would be akin to publishing your private
key to a blockchain... clearly a terrible idea.

In order to do safe biometrics with a blockchain you need:

1. A system (A) that can produce a non-reversible biometric template.
   That is, you can't go from the biometric template to an image of
   the person or anything else that can be re-used to trick the
   system.
2. A system (B) that can check a biometric template against input data
   (image, interactive video stream, etc.).
3. A system (C) that is capable of generating input data to system (B).

System (A) and system (C) can be fully self-sovereign, under the control
of the person represented by the DID. This means that you are also not
handing any of your biometric information over to a 3rd party, you are
in control of your biometrics at any given point in time.

There are protocols that work like what I describe above (and even most
proprietary protocols work in more or less the same way):

https://standards.ieee.org/findstds/standard/2410-2015.html
http://grouper.ieee.org/groups/2410/index.html

In any case, these sorts of biometric templates are safe to put on a
blockchain as long as they have the same qualities as a one-way hash AND
even if they're broken, rotating the template is easy.

-- manu

-- 
Manu Sporny (skype: msporny, twitter: manusporny, G+: +Manu Sporny)
Founder/CEO - Digital Bazaar, Inc.
blog: The State of W3C Web Payments in 2017
http://manu.sporny.org/2017/w3c-web-payments/

Received on Thursday, 4 January 2018 22:18:27 UTC