Re: [Cfrg] I updated 3 drafts related to a FSU KeyEX

One thing worth noting: Multi-authority large-universe ABE (such as [1]) 
entirely subsumes IBE, with some very nice properties that help avoid the 
issues brought up.

In such a system, an authority can act as an IBE KGC trivially, providing 
the same behavior. However, the ability to express audiences in terms of 
monotonic boolean functions of attributes _across multiple authorities_ 
changes the landscape significantly, enabling any of the following:

- Classic IBE use case, "send to user@example.com as issued by 
example.com authority"
- Distributed-PKG IBE, "send to user@example.com as issued by authorities 
A, B, and C" (but here, the sender chooses the authorities to trust, on a 
per-message basis!)
- IBE with pinning (only requires a single additional attribute!), "send 
to user@example.com as asserted by example.com authority, and holding 
'verified by eternaleye@gmail.com' as issued by me"

[1] http://eprint.iacr.org/2015/016

On Thu, 28 Apr 2016 10:26:38 +0100, Stephen Farrell wrote:

> Hi Mike,
> 
> On 28/04/16 09:35, Michael Scott wrote:
>> Maybe the more accurate phrase "n uniquely attractive targets" where
>> n=2,3,4... doesn't carry quite the same punch!
> 
> I'm sorry, but for me, it does have exactly the same punch. If there are
> key generators, they can collude or be coerced. Or even more likely, in
> a realistic commercial Internet-scale deployment, it's quite likely all
> of them (even if operated by different entities) may be running on one
> or two mega-hosting platform,
> so there may well be only one thing to break into even if it looks like
> N things.
> 
> From my POV, the mandatory key escrow aspect of IBE is basically fatal
> for all but possibly some small set of niche applications.
> 
> Cheers,
> S.
> 
> 
> _______________________________________________
> Cfrg mailing list Cfrg@irtf.org
> https://www.irtf.org/mailman/listinfo/cfrg