Re: Fwd: [Jsdsi-users] KeyStore

[Sean R. <sra...@ae...>]

Dav Coleman wrote:

>On Thu, 12 Aug 2004 11:23:22 +0100, Sean Radford
><sra...@ae...> wrote:
>  
>
>Hi Sean,
>
>I have to admit I'm still a little confused. I took a look at the
>CertPathTest and it uses Loader to load some plain text files with
>certificate representations, but they aren't a text format I was
>expecting (the sexp lisp-like stuff), and it's not clear how you would
>create them (there's no Saver class that I found). The
>SexpMarshallingTest source was good for showing me a simpler way of
>creating Name/Auth certs however.
>  
>
Don't worry about being confused. It does take some time I'm afraid, 
mainly because there is currently very little / no help or  getting 
started documentation.

>What is the standard/best/accepted way of creating "your own" keypairs
>(Prinicpals) and saving them to the file system between application
>invocations? And what about storing other people's public keys in your
>name space? I assume in X.509 all this was done in a single KeyStore
>(which comes with load/save methods), but apparently that's not
>correct for a SPKI system?
>  
>
Creating KeyPairs: jsdsi.util.KeyPairFactory.create("RSA", 512);

Saving a KeyPair to file.... For this I think we really need an 
implementation of a KeyStore that supports JSDSI objects (well 
SExpressions really). If I remember correctly the Sun and BouncyCastle 
ones do not work (you can save but they have trouble loading). I did 
have a modified BouncyCastle keystore a long time back that worked(I 
took the original source and added about 3 lines of code) - not sure 
where it is now but could hunt down I guess...

To store other peoples public keys one would create NameCertificates for 
them and store them in a CertStore. If there is a limited number of 
proof searching using a jsdsi.Prover is not required I guess you could 
use a KeyStore?

>The FOO Camp is the one at O'Reilly headquarters in Sebastapol, CA.
>It's in mid september so I figure my chances of finishing this in time
>aren't great (especially with 10 days at Burning Man between now and
>then), but I'm seriously going to try. I have friends going to the
>European one, although I'm not sure when/where that is exactly. If I
>had it done in time they could possibly do a short demo for me there.
>
>  
>
The one in Europe I think is at the end of this month. Somewhere in 
Scandanavia I think. I'll hunt down my modified bouncastle keystore if 
that will help you get going?

>The system I'm trying to build is a trusted brokered p2p network (and
>eventually, a trusted brokered p2p darknet) that uses SPKI/SDSI to
>control authentication/authorization in a way that maps to social
>network concepts. It's just a general concept at the moment, but I
>need to get something to play with to flesh it out. I'm comfortable
>with the network and application programming, but I'm new to PKI
>programming. I've been a PGP user for years but the concepts I've
>learned from that are probably hurting as much as helping.
>
>
>  
>
Stick with SPKI! ;-) (I'm probably biased) Seriously though, I should 
imagine that the A&A stuff in SPKI is what you need, and correct me if 
I'm wrong, but PGP doesn't really support authorisation?

Regards,

Sean

-- 
Dr. Sean Radford, MBBS, MSc
sra...@ae...
http://www.aegeus-technology.com/