[standards-jig] Pondering DTCP

Tom tlburnett at ucdavis.edu
Fri Dec 13 11:06:47 UTC 2002

> Justin Karneges wrote:
> >One company I've been talking with absolutely wants peer-to-peer messaging 
> >capability, to reduce server load.  This would make their service
similar to 
> >ICQ, where thru-server messages are just a fallback.
> >
> Maybe Jabber is then not the right product for them.
> You can't change something because there is one company that doesn't 
> like it. - I think there are many people (me included) that like Jabber 
> because it does not establish direct peer to peer connections. This is 
> additional security for the users and add portability (e.g. what I 
> already have: that an IPv4 user can communicate with an IPv6 user)
> Tot kijk
>     Matthias
I really don't need my server admin sniffing my freaky p0rn.  We all know
Jabber is an open standard and not subject to the whim of you, I, or any
one entity.

IMHO, there's no reason to burden the server with certain types of
interactions.  This is because I come to Jabber from a user perspective and
therefore my bandwidth is limited.  Any server I throw up will be no better
than a DSL connection.  And if I get the people at UC Davis to one day
throw up a campus wide Jabber server, p2p is the way to go.  27,000+
clients could be a bitch.  Do I really need to bother the server when
sending a project file to a group member?

I think there needs to be an option for direct messages.  That's one thing
I think ICQ got right.  It's so elegant.  It's so easy.
You got a port open?  Good.  Here's your message.

I think this is really a server issue and needs to be addressed with a
configuration option.  If security is such an issue (and I understand it
can be), then obviously you won't be using the default install and
configuration of a Jabber server anyway.  You should be able to simply turn
off p2p stuff.  (Like the server won't tell you the IP and port behind the
jid unless it's enabled.  Or in DTCP's case, reject DTCP requests.)

tlburnett at ucdavis.edu

More information about the Standards mailing list