Merit Network
Can't find what you're looking for? Search the Mail Archives.
  About Merit   Services   Network   Resources & Support   Network Research   News   Events   Home

Discussion Communities: Merit Network Email List Archives

North American Network Operators Group

Date Prev | Date Next | Date Index | Thread Index | Author Index | Historical

Re: Digex transparent proxying

  • From: Jay R. Ashworth
  • Date: Wed Jul 01 18:27:13 1998

On Tue, Jun 30, 1998 at 02:24:43PM -0700, Chris A. Icide wrote:
> Against my better judgement, I've decided to add my tiny bit of content to
> this thread.  If there is anyone to thank for getting me into it, Mr.
> Porter has done so with his excellent review of the thread.  Without it, I
> might not have seen the humor in it.

Me too; there were one or two points that I'd like to sppeak to in your
excellent summary, Chris.

> 3. The Dow Jones Industrial Average will never break 4000

Um, I think you misspelled "10,000". :-)

> Events such as the infamous 7007 event, 

Forgive me revealing my ignorance, but what was that?

> Finally, the ethical question comes around.  Is it, or is it not ethical
> for someone to use a transparent cache?  IMHO, the arguments on this
> portion of the thread have been nothing but sensationalist.  It reeks of
> paranoia.  Perhaps the X-Files can (or have they already) done a show on
> this.  THEY are stealing your packets without your knowledge! THEY are
> monitoring your every transmission and receipt!  THEY know your innermost
> secrets!  Bah!

Indeed.  But that wasn't really the argument.  You touch on it next.

> In the case of transparent caching, it's my belief that as long as an end
> user receives content as provided by the content provider in the format
> specified by the content provider (yes, this includes dynamic information),
> both the content provider and the end user could care less what physical,
> electromagnetic, or optical transformations take place from end to end.  If
> the end to end transit is truly transparent, then the goal is accomplished.

This is an example of something I see happen all too often lately.

Any given system has obvious, expected, intended outcomes, and
non-obvious, but yet still desired ones.  When designing a system to
replace an earlier system, the analysis which creates the requirements
document -- the list of things the system must do -- must be thorough
enough to include _all_ the effects of the current system, including
side effects that people have come to take advantage of; not just the
items that were part of the original design specification.

For example: caching.  Over the course of the evolution of the web, the
lifetime of the contents of a page has declined, and the changes in
these pages are not always merely trivial; someone's earlier point about
stock quote pages is pertinent here.  The assumptions about caching
which were valid when the technique was first introduced are becoming
less so, and the limitations of such a technique are becoming more
important to the people at each end.

That this is true is the reason questions of ethics are being raised on
this topic.  I concur with the people who say that "truly transparent"
caching is acceptable... but it's impossible to build a _truly_
transparent cache without a mathematically precise definition -- at the
systems level -- of the protocol involved... which we do not currently
have, nor are we likely to have it any time soon.

OTOH, I also concur with the people who say (this is the majority
opinion I've seen advanced) that the crime Digex is committing is
introducing the proxy (ie: changing the semantics of the protocol)
without notifying any of it's customers that it was, in effect,
_breaking_ the protocol for some of their possible uses, without
telling anyone.

Since I've had bad experiences with Digex's new parent, the 10-years-
and-still-no-profit Intermedia Communications, this policy doesn't
surprise me over much...

> However, the expectation that this goal could be accomplished while content
> providers, transit providers, and client/server software providers work
> independently of each other, is ludicrous.  These entities must work
> together to achieve this.  The pointing of fingers and blaming of the other
> parties does nothing more than delay the maturation process of this
> industry.  This is a complex and extremely interactive system, and must be
> dealt with as a whole for an optimal solution to arise.

Precisely.  Obviously, if there are half a dozen "transparent proxy"
boxen, and only one that peopl seem to think is "truly" transparent,
then we still have a _long_ way to go.

> Perhaps it's the place here for a Grand Statement concerning the likelihood
> of this occurring.  :)

"Always and never are two words you should always remember never to say."

Cheers,
-- jra
-- 
Jay R. Ashworth                                                jra@baylink.com
Member of the Technical Staff             Unsolicited Commercial Emailers Sued
The Suncoast Freenet      "Two words: Darth Doogie."  -- Jason Colby,
Tampa Bay, Florida             on alt.fan.heinlein             +1 813 790 7592

Managing Editor, Top Of The Key sports e-zine ------------ http://www.totk.com




Discussion Communities


About Merit | Services | Network | Resources & Support | Network Research
News | Events | Contact | Site Map | Merit Network Home


Merit Network, Inc.