From: "Chevalier Dev"

>This might change things a bit:
>http://www.infosecurity-magazine.com/view/4997/giesecke-devrient-play-secure-android-card/
>Executive summary: G&D is offering support for a secure element inside Android.

>This might be a good opportunity to gather everybody around a single
>keystore, and from there expand key management to all applications in
>need of it: encrypted e-mail and mass storage, VPN, WiFi, etc. Step by
>step we would end up with an enterprise-level Android.

It is nice to see that somebody is actually doing something in this space!

Since I work with this both as a technologist (looking forward....) and as a 
developer
shipping somewhat more down-to-earth solutions (...), I'm definitely interested.

IMO, the goal (of any such endeavor) should be to make it a de-facto standard.
To do that the entire scheme including the SE must be published.   That a card 
is
based on GlobalPlatform or 7816 etc. have in practice not returned any 
realizable
interoperability and upgrading phones is much more difficult than PCs with 
Windows.

Now to the applications.  SEEK seems to be oriented towards enterprises.  VPNs
using IPSEC or SSL ought to be the killer apps because it applies to the entire 
spectrum
of interconnected services.

My personal take on this is that Google should (as the consumer-company they 
are),
put more efforts in consumer solutions, which if properly designed, should be 
quite
useful for enterprises as well.

What's the problem here?  Basically it is that consumers means browsers and the 
web.
Unfortunately Google and the Android team appear unaware of the fact that 
Netscape's
13-year old relic known as <keygen> simply has nothing to offer smart card 
solutions
like SEEK.

OTOH, I'm not completely convinced that SEEK actually matches consumer needs 
either
because the GlobalPlatform consortium haven't AFAICT spilt much energy on *open*
on-line provisioning which is the most likely consumer method for 
authentication keys.

The following is a scheme in early stages that tries to clone SEEK with 
consumer needs:
http://webpki.org/papers/keygen2/secure-key-store.pdf

Although the details of SEEK are sketchy I *assume* that the main differences 
between
SKS and SEEK are that the former :
- Supports air-tight key-provisioning and key-management
- Is heavy tied to a protocol called KeyGen2
- Reserves some of the FLASH memory for storing certificates giving
  room for hundreds of keys including Microsoft InformationCards
- Does not support Java, 7816, or PKCS #15, only a cryptographic API

Regards
Anders Rundgren


Reply via email to