ietf-openproxy
[Top] [All Lists]

Re: IAB Requirement for OPES

2001-11-26 11:06:48

Sorry, this is a late reply to Markus's question about intermediaries
that only log/bill.  (I didn't realize until now that this didn't
make it to the mailing list, because I was not a member...)  

- Sally

------- Forwarded Message

...
Message-Id: <200111192032(_dot_)fAJKWjr14941(_at_)elk(_dot_)aciri(_dot_)org>
To: Markus Hofmann <hofmann(_at_)bell-labs(_dot_)com>
cc: OPES Group <ietf-openproxy(_at_)imc(_dot_)org>, 
leslie(_at_)thinkingcat(_dot_)com
From: Sally Floyd <floyd(_at_)aciri(_dot_)org>
Subject: Re: IAB Requirement for OPES
Date: Mon, 19 Nov 2001 12:32:45 -0800
Sender: floyd(_at_)aciri(_dot_)org

I'm just wondering whether this consideration rules out realizing such
services in an OPES framework and whether this is intentional or
whether we need more clarification on this.

Well, that item in the draft was mostly motivated by robustness
issues of dealing with broken/imperfect/hacked OPES intermediaries
that modify data, but it did say that some of the same issues, of
robustness with intermediaries that *could* modify data but aren't
supposed to, apply to plain vanilla web caches as well.

So my own hit would be that for intermediaries that only log/bill,
the question has to be asked of how the network could be robust
in the presence of failures.  I don't have a vote for the answer,
and guess that many such issues will come up, that will have to
be dealt with by the OPES group, and confincing agruments resented
                                      convincing arguments presented
one way or another. 

Sorry, but I am late for my meeting in Tucson...

- Sally

...

------- End of Forwarded Message


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