Re: [JXTA discuss] NEW Re: [JXTA discuss] Mobile IP

Mohamed,

Thank you very much for your answer. A couple of new questions came up:

1) How the APA is created and managed? Internally fromt he JXTA impl? Is
there a daemon in each peer that republishes an APA advertisement every time
the available endpoints changes?

2) What happens exactly when we have a relay peer? The edge peer that
connects through the relay polls the relay in intervals for any new
messages? Or the relay caches messages for the edge peer until he finds out
the new APA.

3)I assume the 3-message exception you mention appears on the relay or the
peer that tries to connect directly to the mobile peer.

4) If we have a multihomed peer that has more than one different network
interfaces (Bluetooth,WLAN, etc.) is there an Endpoint for each interface or
one Endpoint for all? How does JXTA choose the network inteface [
Endpoint(s)]  through which the communication occurs?

Thank you very much for helping me in a deeper understanding of how JXTA
works.

Best regards,

Giorgis Georgakoudis

----- Original Message -----
From: "mohamed abdelaziz" <[email protected]>
To: <[email protected]>
Sent: Tuesday, July 19, 2005 1:21 AM
Subject: Re: [JXTA discuss] NEW Re: [JXTA discuss] Mobile IP


> Giorgos,
>
> A peer (jxme, or otherwise) can be mobile about the network, without
> having to worry about it's new address. Addressing within the JXTA
> network is by a peer's endpoint, which is resolved to an access point
> advertisement (an APA describes a peer's physical and virtual
> endpoints), if a peer changes it's physical or virtual endpoint, such
> change is reflected within the access point advertisement, and is
> automatically re-resolved when connection failures occur.   In relayed
> situations where the relay remains the same, a change in physical
> addresses has no effect, whereas in direct connections, or when a change
> to a different relay occurs, such change would manifest itself as slower
> than usual delivery of messages to the mobile peer, or in an exception
> when attempting to send more than 3 messages before the new Access Point
> Advertisement is resolved.  Also keep in mind that peer routes are
> resolved by a peer's advertisement or access point advertisement (via
> discovery), through the EndpointRouter resolution protocol, of through
> forward routes which are part of delivered messages.
>
>
> Cheers,
> Mohamed
>
>
>
>
> Giorgos Georgakoudis wrote:
>
> >I read the paper but I did not get any clear answer. Ok, here is a
scenario:
> >
> >Someone is using a JXTA enabled PDA communicating in P2P JXTA network.
> >Moving in the city, the PDA's IP changes from X to Y. What happens?
> >
> >Guess 1: The Peer Advertisement that the PDA has issued still exists (as
the
> >timeout has not expired) so when other peers want to communicate with
this
> >peer they find a "new" route. There is some delay though. ( Will there be
> >any missing packets granted they were routed according to the "old"
route?
> >Is there any caching of messages in the relay route?)
> >
> >Guess 2: The peer is considered disconnected so it issues a new Peer
> >Advertisement to manifest its existence in the JXTA universe. All
previous
> >communication is considered disconnected.
> >
> >Giorgis Georgakoudis
> >
> >----- Original Message -----
> >From: "Peter Fagerlund" <[email protected]>
> >To: <[email protected]>
> >Sent: Saturday, July 16, 2005 3:14 PM
> >Subject: Re: [JXTA discuss] Mobile IP
> >
> >
> >It do not affect the jxta protocols, the mobile is recognised through a
> >jxme-proxy relay being the window in to the jxta world.
> >
> >http://www.jxta.org/project/www/docs/JXTA4J2ME.pdf
> >
> >/p
> >
> >On 16 Jul 2005, at 10:24, Giorgos Georgakoudis wrote:
> >
> >
> >
> >>Hello,
> >>
> >>
> >>I would like to ask what happens in jxta when we have a mobile peer
> >>that is moving through networks thus having a mobile IP address. How
> >>does that affect the jxta protocols? Is the peer considered dead and
> >>reconnected or otherwise? If there is any documentation mentioning
> >>this or any other concepts please let me know.
> >>
> >>Giorgis Georgakoudis
> >>
> >>
> >
> >
> >---------------------------------------------------------------------
> >To unsubscribe, e-mail: [email protected]
> >For additional commands, e-mail: [email protected]
> >
> >
> >
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [email protected]
> For additional commands, e-mail: [email protected]