ietf-openproxy
[Top] [All Lists]

RE: Tentative OPES Agenda for IETF 53

2002-03-12 15:47:03
Hello Ian,

comments inline...

-----Original Message-----
From: Ian Cooper [mailto:ian(_at_)the-coopers(_dot_)org]
Sent: Tuesday, March 12, 2002 11:55 AM
To: OPES Group
Cc: Penno, Reinaldo [SC9:T327:EXCH]
Subject: RE: Tentative OPES Agenda for IETF 53


--On Tuesday, March 12, 2002 10:58 -0800 Reinaldo Penno 
<reinaldo_penno(_at_)nortelnetworks(_dot_)com> wrote:

I second John's email. Personalization has a lot to do with OPES.
Personalization in relation to OPES have a lot of interested 
parties as
you can see based on the two drafts we produced. You can say 
it's not on
the immediate charter but it is something we should pursue without a
doubt.

If it's not on the charter then it's not the primary goal, per 
se, of the 
group.  Sure it's something that has to be considered but I 
think it should 
be viewed as one of the services that can (should) be provided 
within the 
OPES architecture.

What is going to be I'm not sure, but clearly is not in the charter but has
a lot of interest.


Moreover, when you talk about endpoint authorization, the everlasting
question (that we discussed ad nauseum in PANA)  is: are you talking
about the device or the user? In the case of OPES you are 
surely talking
about the user, since the authorization needs to be given on 
a per user
basis.

Be careful about the term "user".  I think you mean 
User-agent, or more 
possibly Client in the HTTP-land.  User and device are 
different entities 
and may require their own independent authorizations, surely.

agreed, that was what I said...authorization is good only for the
User->User-Agent->Device that in that moment in time is using the service.
What happens if two users share the same PC and one has content filtering
and another has not? What happens if a machine has more than one interface?
If the user is roaming? IMHO A lot of device/user identification can be
taken from the PANA work.  

Also, are you going to authorize that every single page is filtered?  


In this case you can say that OPES is by definition a personalized
service.

Personalization is one of the things it gets you but I think 
the term is 
too loaded to be useful.  Is the (ever present) virus scanning 
example a 
case of "personalization"?  (This feels like a rat hole, so let's be 
careful)

yes, but I would say its a weak example. Some user might have some not.
Different users that sharing the same PC, one has virus scanning another
not. If a user roams to another network, he would like stil be able to use
its virus scanning service.





<Prev in Thread] Current Thread [Next in Thread>