[prev in list] [next in list] [prev in thread] [next in thread] 

List:       jabber-jdev
Subject:    [jdev] XEP174 with DHT? (serverless jabber)
From:       "Michael Schmidt" <schmidtm524 () googlemail ! com>
Date:       2007-10-23 15:07:19
Message-ID: 5cc93d950710230807x32a77483p930f529f50c3a7b1 () mail ! gmail ! com
[Download RAW message or body]

Hi,

there was a concretization, how jabber goes serverless with the jabber protocol
and how to discover the other users. Maybe a DHT should be used, which
is still now not
 foreseen in the description of XEP 174?  So Hal Rottenburg from the
jabber client list suggested to get help on this jabberlist.

If you make it concret for a client (here: PSi discussion), how is
then a user discovered for XEP 174? Which DHT should be used, will
there be one created as Jabber DHT?
Read the added text from bottum->up to maybe help this question.

Thanks Regards Mike


2007/10/23, dev <dev.akhawe>:
> Hey,
> point is ... PGP OpenPGPE / SSL can be used as e2e encryption if it is serverless.
> But PGP  can't be used if you are routing it through the server. and
> SSL kicks ass unlike some of the other encryption proposals we know :)
> .. It is simple , fast and well tested .
> see the development here: http://retroshare.wiki.sourceforge.net/devel-core-openpgp
> I don't know how valid this is for XMPP(meaning if it is possible to
> use SSL in serverless XMPP) , but this is what I have noticed in
> general for any distributed system
>
> my 2c.
>
> Regards,
> dev
> _______________________________________________
> Psi-Devel mailing list
> Psi-Devel@lists.affinix.com
>
>


2007/10/23, Hal Rottenberg <hal>:
> > A DHT is much better than DYNDNS or forholding the IP, which may dynamically.
> We are major
> proponents of the XMPP protocol and what it stands for.  Arguments for
> DHT are irrelevant as there is not a peer-reviewed,
> standards-body-approved, method of using distributed hash tables for
> IM and presence that has anything to do with XMPP.
>
> If you have a compelling technical argument to make, make it on JDEV
> (and bring your own implementation because that's a developer list) or
> the XMPP Protocol Extension mailing lists.  Those can be found here:
> http://www.xmpp.org/about/discuss.shtml
>
>
> --
> Blog: http://halr9000.com
> Webmaster, Psi (http://psi-im.org)
> Co-host, PowerScripting Podcast (http://powerscripting.net)
> _______________________________________________
> Psi-Devel mailing list
> Psi-Devel@lists.affinix.com
>
>



2007/10/23, Andreas Ntaflos <daff>:
> On Tuesday 23 October 2007 15:08:00 Remko Tronçon wrote:
> > > A DHT is much better than DYNDNS or forholding the IP, which may
> > > dynamically. So lets talk about the serverless implementation into PSI.
> > > Why not a DHT?
> >
> > What is DHT? And DynDns is indeed useless, since requires a central
> > (DNS) server, so you're no longer 'serverless'.
>
> DHT means, as far as I know, Distributed Hash Table and can be used by
> peers/nodes in a distributed system to talk to each other without a central
> authority/server managing the communication. Trackerless Bittorrent is a
> popular use case for this I think.
>
> > Why do you think serverless messaging is so interesting?
>
> Serverless IM probably allows
> more secure communication between two people as there is no server in-between
> them that could have a malicious admin who likes to eavesdrop or do a
> man-in-the-middle-attack. Of course using GnuPG for end-to-end encryption
> would effectively prevent such a scenario but knowing how most people think
> setting up GnuPG is too big of a hassle.
>
> Serverless IM presumably would feature encryption on the transport level
> somehow so it woud be easier and more secure "out of the box", without users
> having to set up end-to-end encryption manually.
>
> Assuming, of course, I understood the concept correctly.
>
> Andreas
> --
>
> _______________________________________________
> Psi-Devel mailing list
> Psi-Devel@lists.affinix.com
>
>
>


2007/10/23, Remko Tronçon <remko>:
> > A DHT is much better than DYNDNS or forholding the IP, which may dynamically.
> > So lets talk about the serverless implementation into PSI.
> > Why not a DHT?
>
> What is DHT? And DynDns is indeed useless, since requires a central
> (DNS) server, so you're no longer 'serverless'.
>


> Kevin, thanks for the info,
> this was a real interest, heard about jabber serverless, but did not
> know, that is is already such elaborated, that clients build it in. As
> I read here, for the RS DHT a protocol adjustment is planned and
> possible, maybe they switch as well to xep174.
> http://retroshare.wiki.sourceforge.net/devel-core-im
> Would you suggest in this development?
> Did you read this:
>
> "4. Discovering Other Users
>
> In order to discover other users, a client sends an mDNS request for
> PTR records that match "_presence._tcp.local.". The client then
> receives replies from all machines that advertise support for
> link-local messaging. [11] The client MAY then find out detailed
> information about each machine by sending SRV and TXT queries to
> "username@machine-name._presence._tcp.local." for each machine
> (however, to preserve bandwidth, the client SHOULD NOT send these
> queries unless it is about to initiate communications with the other
> user, and it MUST cancel the queries after it has received a
> response). Note: The presence name to be used for display in a
> link-local "roster" SHOULD be obtained from the <Instance> portion of
> the received PTR record for each user; however, the client MAY instead
> display a name or nickname derived from the TXT records if available."
>
>
> A DHT is much better than DYNDNS or forholding the IP, which may dynamically.
> So lets talk about the serverless implementation into PSI.
> Why not a DHT?
>

> 2007/10/23, Remko Tronçon <remko>:
> > The lack of implementations of other protocols is compensated by the
> > fact that Jabber allows you to connect to other networks server-side
> > (transports),
>
> 2007/10/23, Kevin Smith <kevin>:
> > If There is a Jabber protocol for serverless messaging, and Psi will
> > implement it.
> >
>
>
> serverless is better than serverbounded.
> A Transport is always serverbounded (ok as the original protocol as well,
> but you need to register a second server account in jabber to get access to aol)
>
> so serverless jabber is quite interesting
> found this jabber2 description:
> https://sourceforge.net/tracker/index.php?func=detail&aid=1772358&group_id=178712&atid=886242
>


2007/10/23, Kevin Smith <kevin>:
>
> If There is a Jabber protocol for serverless messaging, and Psi will
> implement it.
>
>
> The official protocol for serverless messaging is available at
> http://www.xmpp.org/extensions/xep-0174.html
> We will support this in the future.
>
>
> /K
> _______________________________________________
> Psi-Devel mailing list
> Psi-Devel@lists.affinix.com
>
>

[prev in list] [next in list] [prev in thread] [next in thread]