Web Informant #285, 26 April 2002:
Learning from my students on network security products
As a beginning high school teacher, this year I have found
that you have to take advantage of those "teaching moments."
These are unfortunately (at least, for me) those rare times
when you actually penetrate your students' minds and get
across some significant but hard to understand piece of
information that can explain something you were trying to
dance around for hours, days, or even weeks earlier.
Sometimes it is called an "aha" -- as when the lights go on
and the student suddenly sits up and pays attention to what
you are saying. I look forward to these moments: They make
the whole teaching experience worthwhile for me.
Well, we teachers have our moments too, when we finally learn
something from our students. And the biggest moments for me
have been dealing with the failure of a couple of network
security products. Let me explain.
I began my school year with many of my kids clamoring to try
to hack into their own networks, asking me to help them use
the same kind of tools that the "real" hackers run every day
out on the big, bad Internet. So, latching on their interest,
I tried to accommodate them, showing them how to use
Ethereal, nmap, netcat and the like, but that didn't work. My
kids didn't want to take the time to muck around with a bunch
of obscure command-line parameters and watch these port
scanners work their way through ten thousand ports and
zillions of IP addresses: They wanted to sniff out their
friends' (and enemies') network and AOL passwords.
So it is somewhat ironic that while I began this year testing
network security, the moments that I have gotten my own
"ahas" have been in the same area. The twist is that I have
learned more from the failures of products than from the
usual hacker tools.
The security failures that I observed were much more prosaic
and have close parallels to the failures of these products
that you would observe in the ordinary workplace. I am
talking about establishing encrypted e-mail and virtual
private networks (VPN) tunnels. Both are products that are
very desirable but are still very hard to implement.
Last fall I took a few of the students aside and had them try
to get Pretty Good Privacy working among us to encrypt our
e-mail traffic. Well, the project fell flat on its face. A few
of the kids got PGP installed on their home computers. But
then when they tried to transfer the keys to the school
computers, they ran into problems. The whole key
infrastructure thing bogged down, and I never could find the
time to debug things or set it up properly to really get
everything working. It might have been our school's firewall,
or the way we lock down the lab computers using Fortress to
keep the kids from messing up their configuration. It might
have been me, even though I have used PGP in the past
successfully.
The real reason, though, is this: We could communicate just
fine with unencrypted e-mail (which is how I send their
homework assignments out), so why bother with all the
security anyway? As I said, this mirrors the state of
encrypted e-mail in the corporate world all too well.
My latest teaching moment was this spring. I have been trying
to set up a VPN between my home network and a student's.
Again, another failure. The idea here was to have each
student be able to view a bunch of shared documents on my
home network, just as they would connect to a local file
server, but across the Internet. That is the nice thing of
having a VPN and something that is desirable in corporate
applications, as well.
To make things easier, we began this experiment by using a
matched pair of Linksys EtherFast model BEFVP41 routers. They
are remarkably easy to install: You just plug it into your
cable modem and attach your computers to its switched hub
ports. It is remarkably hard to configure properly, which is
done with a series of Web-based screens. When I saw the
screens, I thought that my students would take to them like
ducks in water. No fussy and obscure command line parameters!
Finally, something that could speak their language. Well, I
was partly right.
But so far two of my better students have brought home the
routers and we've had no luck getting connected to each
other's home networks. I know the product works: I spent an
hour with one of Linksys' technical support folks and we got
our tunnel up and running just fine. But the product isn't
really ready for corporate use or even student use. It isn't
because my kids aren't sharp; they are. It isn't because they
aren't motivated to tinker with the product and get it
working; they tinker with tons of stuff on their computers
every night, probably too much to the concern of their
parents. It is just that the Linksys routers have too many
knobs to tweak and to turn before they will work properly.
You might want to spend some time pre-configuring them before
you ship them out to the hinterlands.
Network security products like PGP and low-end VPN routers
have their place, to be sure. But they are still way too hard
to use for the average person, even a highly motivated
teenaged geek.
Entire contents copyright 2002 by David Strom, Inc.
David Strom, david@strom.com, +1 (516) 944-3407
Port Washington NY 11050
Web Informant is (r) registered trademark with the
U.S. Patent and Trademark Office.
ISSN #1524-6353 registered with U.S. Library of Congress.