RE: Private WiFi in hotel room; security/interference issues

Tim --

I would really count on using PC-Cards, since they can be had in a/b/g
configuration for maximum flexibility (though cards having "a" tend to be
much more expensive than the commodity b/g cards).  The evolution of
wireless is still going on at a furious pace, so you may well want to
upgrade in the future. And personally, I would not compromise my selection
of laptops to accommodate some built-in wireless capability.  Are there so
many crooks in the Perl/UNIX community that theft of cards is a real
problem?  Or, of course, you can check-out the cards -- and laptops -- to
the students and make them accountable.  Just that might thwart any
temptations.

It is my understanding that "a" and "g", at least "standard g", have similar
data rates, so there would be no advantage to "g", except range and generic
availability; and "a" also provides more channels.  Note that "a" channels
are split between two regulatory regimes, as I understand it.  One really
demands a built-in, "pre-qualified" antenna, whereas the other is more
flexible (though, theoretically, only radio/antenna combinations that have
been tested and certified are truly legal in any mode, but we'd need to
double the FCC's staff to get a bead on that one).

Beware of semi-proprietary implementations of advanced protocols whose specs
are still in IE3 draft form.  Manufacturers try to get a jump on the field
with implementations of the draft and filling-in the blanks themselves, but
when the draft is finally blessed, it is possible that these implementations
will become orphans.  Of course, if you are using compatible AP's and cards,
it doesn't matter until you want to begin upgrades, or when someone wants to
use his own laptop in class.  However, if there is fear of theft, then
explaining that you are using a proprietary solution may cool them off a
bit.

Perhaps someone on this list can give advice on antennae, especially for
"a", as it seems to have become the standard position that a directional
antenna is much more valuable to optimizing performance than even adding
reasonable amounts of power would be.  Since antenna patterns can limit
vertical dispersions as well, this would help to keep the signal out of the
guest rooms.

But I still don't understand this paranoia.  There may be better security
modes than the standard but eminently crackable WEP available for the AP's
and cards that you select.  However, WEP is still the security equivalent of
locking your car -- determined thieves hardly notice, but casual temptations
are suppressed.  But who cares anyway?  As I said, Perl is almost an
encrypted language itself :-)  And remember, you can suppress the broadcast
of the wireless net ID's, so only your students will know that your network
exists -- it won't show up on just anyone's available list -- security by
obscurity.

Ken Meyer


-----Original Message-----

From: [email protected]
[mailto:[email protected]]
On Behalf Of Tim Maher
Sent: Tuesday, January 10, 2006 9:03 AM
To: [email protected]

Subject: Private WiFi in hotel room; security/interference issues

Greetings!  I've recently heard about seattlewireless.net, and
I've read the FAQ, but I still need some help in deciding how to
configure a temporary, autonomous WiFi network in a hotel.

BACKGROUND: Wired to Wireless Transition

While running my periodic computer classes in area hotels for the
last 15 years, I've set-up and torn-down enough temporary wiring
to last me the rest of my life. Not to mention all the duct-tape
trees I've killed in the process! 8-}

So now I'm looking into the possibility of trading in my old
hardware configuration for a new one, based on laptops and
servers using WiFi connections.

I'm contemplating this move because I'm thinking WiFi technology
might be good enough by now to provide reliable service in a room
of 750 sq ft or so, with all client machines being within a
distance of 30 feet, and having line-of-sight (or nearly so) to
the access point.

I don't know much about WiFi ('cuz I've favored wires), so I'm
appealing to the combined wisdom of the WiFi wizards of this list
for some pointers on my proposed setup.

THE NETWORK

I'll be running a server equipped with an "access point" in
various hotel meeting rooms, with up to 16 WiFi- equipped laptops
(provided by me) as clients. In some cases, Internet access may
be available from a wall-plug to the server machine, which could
be routed to the clients. But hooking up the client laptops to
the server through WiFi is the more important issue.

QUESTIONS

1) Which 802.11 protocol would be best for this application?

To minimize interference with the hotel's own WiFi service, which
would undoubtedly be in the 2.4Ghz band, I'm wondering if I
should consider using 802.11/a.

If I configured a /a-only network, I understand that the range
would be limited to 30ft (although some other sources say 100
feet), but either one would be sufficient to cover the classroom.

I might also benefit from using /a to avoid interference with the
hotel's wireless phones and microwaves (which could in some cases
be as close as on the other side of the meeting-room wall).

But I must admit, I don't quite understand the implications of
interference; does it translate into dropped packets, or
retransmissions, or what?

On the other hand, although usability is more important to me
than state-of-the- art status, within reasonable limits, I'd
certainly rather have the throughput of /b or /g rather than /a
if possible.

BTW, obtaining /a network cards would be doable, because I'm
planning to buy reconditioned older laptops, which are available
in /a and /a/b configurations, as well as b/g configurations.
(For theft-prevention reasons, I'm planning on using built-in
WiFi, not PCMCIA cards, so I need to decide on the protocol
before deciding on the laptops.) But the server machine wouldn't
be a laptop, so I'd have to be able to find a card (/a,
especially) for that one.

So should I consider /a, or is /b/g likely to be satisfactory?

2) Can the multi-protocol cards be forced to use a particular
   protocol? If so, then there'd be no advantage to having an /a-
   only or /b-only card over an /a/b or tri-mode card, right?

3) Are wireless routers still available for the /a protocol? I'm
   no stranger to RE-PC, but I've never looked for such a thing
   there yet.

4)
   a. Is WiFi security well enough developed by now so that I
   could be confident that a LinkSys WRT54GS router in some
   standard (or reasonably achievable) configuration (if I go
   with the 2.4Ghz band) would keep hotel guests off the systems
   in my network?

   b. Should I use static IP to further discourage potential
      intruders? (Remember, this isn't a coffee-shop network;
      I'll own all the client laptops, and be free to set their
      IP addresses.)

   c. Would there be an advantage to setting this network up as
      point-to-point, and just using the Linux firewall tools to
      restrict student-to-student "port weaseling" and such?

5) For Linux, are there particular mini-PCI WiFi cards
   that should be preferred, or avoided? (I know about the Linux
   Hardware List, but it concentrates on what cards /have/ drivers,
   not which driver/card combos /work best/.)

   Most of the laptops I'm considering come with Intel cards, but
   there are a few other choices. Here are some of the names:
     * Intel Pro/Wireless LAN with Intersil Prism 2.5 chipset
     * Intel Pro/Wireless 2100
     * Cisco Aironet Wireless 802.11b
     * Atheros AR5001X chipset (11 a/b)

   I've seen postings about these following cards--which are
   available in some other laptops I'm considering--having
   unsatisfactory drivers, but I can't tell if that problem's
   been fixed or not by now; can somebody tell me?
     * Intel Pro/Wireless 2200BG
     * Intel Pro/Wireless 2915ABG

6) Of course, WiFi cards aren't expensive to replace, so if
   somebody has a favorite one that I haven't mentioned here,
   that Linux likes, I'd be willing to consider that one too!
   But I've heard that replacing mini-PCI WiFi cards can be
   delicate surgery; is this true? FWIW, I'm no stranger to the
   innards of laptops, having stripped several down to their
   motherboards in the past.

7) Or should I just use "good old wires", and stop thinking about
   relying on this fashionable, but possibly unripe technology? 8-}

TIA,
*-------------------------------------------------------------------*
|  Tim Maher, PhD  (206) 781-UNIX   (866) DOC-PERL  (866) DOC-UNIX  |
|  tim at ( Consultix-Inc, TeachMePerl, or TeachMeUnix ) dot Com    |
*-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-*
|  Watch for my upcoming book: "Minimal Perl for UNIX/Linux People" |
|  See MinimalPerl.com for details, ordering, and email-list signup |
*-------------------------------------------------------------------*

_______________________________________________
Talk mailing list
[email protected]
http://seattlewireless.net/mailman/listinfo/talk


_______________________________________________
Talk mailing list
[email protected]
http://seattlewireless.net/mailman/listinfo/talk