[11421] in Commercialization & Privatization of the Internet
Re: Re[2]: Two-way Internet service from Continental Cable?
daemon@ATHENA.MIT.EDU (Huon N. Dang)
Fri Apr 1 14:19:03 1994
Date: Fri, 1 Apr 1994 07:33:03 +0800
From: hnd@chevron.com (Huon N. Dang)
To: 0003858921@mcimail.com, Mark-Ludwig@uai.com
Cc: steve@nsf.gov, com-priv@psi.com
Can someone out there please remove my ID (hnd@chevron.com) from this
mailing list. Thanks.
-Huon
> From com-priv-forw@lists.psi.com Thu Mar 31 16:54 PST 1994
> From: "Mark R. Ludwig" <Mark-Ludwig@uai.com>
> To: "Robert G. Moskowitz" <0003858921@mcimail.com>
> Cc: Stephen Wolff <steve@nsf.gov>, com priv <com-priv@psi.com>
> Subject: Re: Re[2]: Two-way Internet service from Continental Cable?
> Date: Thu, 31 Mar 1994 10:52:23 -0800
>
> I'm surprised the more regular contributors who know this better than
> I haven't challenged this misleading impression.
>
> >>>>> "Bob" == Robert G. Moskowitz <0003858921@mcimail.com> writes:
>
> S said:
>
> >If it's a shared medium, how difficult is it for a moderately capable one
> >of PSI's customers on the cable to listen in on what some other customers
> >are saying? Presumably it's slightly more difficult than hanging a
> >Sniffer on a vampire tap, but I wonder what safety mechanisms PSI has to
> >prevent their customers from wiretapping each other?
>
> Bob> Steve, it is soooo simple to turn a PC into an network analyser. I do it
> Bob> all the time at work. A number of commercial packages for ~$1K or PD
> Bob> packages do it...
>
> Yes, and the data vulnerable to this is highly segmented, due to
> network performance requirements. That is, as the network is
> segmented (subnetted), typically for performance reasons, so too the
> ability to sniff is diminished. If you subnet for security reasons,
> you probably get a performance boost as a side-effect. Neat, huh?
> (The exception comes when the majority of the traffic on the original
> network now has to cross from one to the other; then you've gained
> neither for security nor performance.)
>
> I can sniff any packet to or from our network, and that's the end of
> my ability. The same is true most everywhere else in the world that
> there are wires or workstations in cubicles, at least without extra
> effort and/or hardware. I cannot possibly sniff anything you're
> doing, unless you're doing something with our network. On a broader
> scale, the Internet is distributed such that it's just not possible
> without _lots_ of physical connections to sniff any arbitrary person's
> network traffic. You can only sniff what's traveling over the wire
> where you're connected. You want to watch what's happening between
> Europe and Japan? You have to figure out all the ways the packets can
> get from one to the other, and have physical taps on each of those
> wires.
>
> We are a PSI customer. The Internet architecture simply gives us no
> way to wiretap anything which does not involve us, including anything
> to or from another PSI customer. There's nothing special about PSI,
> it's just the way the Internet works.$$
> --
> INET: Mark-Ludwig@UAI.COM NIC: ML255 ICBM: USA; Lower Left Coast
> "For crabby middle-aged movie critics, it was a pretty long night."
> -- Jim Schweda reviewing _D2:_The_Mighty_Ducks_
>
