[JDEV] Routes, etc..

[Waleed Abdulla]

Thomas,
    That's good reasoning. I guess I agree with that. I just feel sorry for
the server developers :)

waleed


-----Original Message-----
From: Thomas Charron <tcharron at ductape.net>
To: jdev at jabber.org <jdev at jabber.org>
Subject: Re: [JDEV] Routes, etc..


>Quoting Waleed Abdulla <WaleedKA at emirates.net.ae>:
>> In fact I think things are getting out of hand. Maybe I'm missing
something
>> here!! Hope, someone can clarify it to me.    DO WE REALLY NEED THE
ROUTE?
>
>  Yes.  It solves an immediate problem that we have, and takes care of some
>longer term issues.
>
>>      I mean, all these tags and attributes added with each message will
>> only
>> make things complex for the servers. I know route info is useful in many
>> cases, but it shouldn't be sent with each message. It should only be sent
>> in
>> a special type of message used to analyze the route (this is how it's
done
>> in IP).
>
>  This data will also be able to be used to handle spoofing issues, which,
IMHO
>are a major concern.  Once we antispood enable the systems, the route
tracing
>data will be needed.  Now, the client doesn't need to SUPPORT it, and the
>clients don;t even have to read the tag if they don't want to, so it's
really
>not expanding the complexity of the client.  Unless, of course, they want
to
>send timezone data along with the message.. ;-P
>
>>      Also, I don't see how the route info can prevent spam. You can't
>> filter
>> all messages coming through route X just because someone using that route
>> is
>> sending spam.
>
>  Becouse we can perform reverse lookups on the actual trace.  We could, if
we
>wanted to make things 100% secure, simply ask a transport if they did
indeed
>send this message, and, in a way, validate the data.  We could also
implement a
>signature that the servers could use that could validate the routing
entries..
>
>>     I vote for simplicity: The sender's client stamps the message and
>> that's
>> it.
>
>  That's all the client needs to do.  This is all server enabling stuff,
and,
>network enabling stuff.  The clients CAN just ignore the tag, and the
sending
>client merely timestamps a node entry within the route tag..
>
>>     Did I say something wrong? :)
>
>  Nope, it's a good point that we not complicate the protocol, but in order
for
>Jabber to become an industrial strength communications protocol, we need
>routing headers, just like email does..
>
>---
>Thomas Charron
><< Wanted: One decent sig >>
><< Preferably litle used  >>
><< and stored in garage.  ?>>
>
>_______________________________________________
>jdev mailing list
>jdev at jabber.org
>http://mailman.jabber.org/listinfo/jdev