[aerogear-dev] iOS Passphrases for UPS

[Bruno Oliveira]

Nailed it, Sebastien.

-- 
abstractj

On March 10, 2014 at 6:12:46 PM, Sebastien Blanc (scm.blanc at gmail.com) wrote:
> Let me try to answer your questions so I can see if I have understand it
> well and then Abstractj can confirm (or not ) :)
> 
> 
> On Mon, Mar 10, 2014 at 10:01 PM, Matthias Wessendorf wrote:
> 
> > Ahoy!
> >
> >
> > On Mon, Mar 10, 2014 at 9:49 PM, Bruno Oliveira wrote:
> >
> >> Good morning guys, for this issue
> >> https://issues.jboss.org/browse/AGPUSH-358. I was revisiting the whole
> >> UPS code and thinking about include two fields for iOSVariant class: skey
> >> (secret key) and pKey (public key). What's the idea?
> >>
> >> 1. Each application has its own key pair
> >>
> >
> > each of the PushApplication constructs/objects on the UPS ?
> >
> Yes, What I understand : each PushAPP has his own pair
> 
> >
> >
> >> 2. Before the addition of the iOS variant, the client sends a request
> >> asking for the public keys
> >>
> >
> > "the client" ? Is that the AdminUI (or the relevant HTTP Rest code), when
> > someone clicks "Add new Variant" ?
> >
> I understand the AdminUI/Rest Service. And this mean there will be an extra
> REST service to retrieve the public key. Idea : maybe the public key could
> be already in the response of the GET of the "Parent" PushApplication.
> 
> >
> >
> >
> >> 3. The server sends an HTTP response with the public key for encryption
> >> 4. The client make use of the public key to encrypt the certificate + the
> >> passphrase
> >> 5. Server stores it encrypted
> >> 6. When necessary to send push messages, the server make use of the
> >> private key to decrypt that data and send fancy messages.
> >>
> >
> > Now, when a (JavaEE) application is sending a request for push messages to
> > the UPS (which than internally makes use of the private key to decrypt that
> > data and send fancy messages to Apple), the HTTP RestEndpoint requires a
> > new argument (the public key) to be allowed to submit these "requests" ?
> >
> I don't think so since it will use only the private key to decrypt, so the
> API stays unchanged.
> 
> >
> >
> >>
> >> Does it make sense to you?
> >>
> >
> > Somewhat - not sure I fully understood :)
> >
> >
> >>
> >> --
> >> abstractj
> >>
> >> JBoss, a division of Red Hat
> >>
> >>
> >>
> >> _______________________________________________
> >> aerogear-dev mailing list
> >> aerogear-dev at lists.jboss.org
> >> https://lists.jboss.org/mailman/listinfo/aerogear-dev
> >
> >
> >
> >
> > --
> > Matthias Wessendorf
> >
> > blog: http://matthiaswessendorf.wordpress.com/
> > sessions: http://www.slideshare.net/mwessendorf
> > twitter: http://twitter.com/mwessendorf
> >
> > _______________________________________________
> > aerogear-dev mailing list
> > aerogear-dev at lists.jboss.org
> > https://lists.jboss.org/mailman/listinfo/aerogear-dev
> >
> _______________________________________________
> aerogear-dev mailing list
> aerogear-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/aerogear-dev