[[IP] New flaw takes Wi-Fi off the air]

New flaw takes Wi-Fi off the air
By Patrick Gray, Security Focus (drew.cullen@theregister.co.uk)
Published Thursday 13th May 2004 21:29 GMT
<http://www.theregister.co.uk/2004/05/13/wifi_security_flaw/>

A newly-discovered vulnerability in the 802.11 wireless standard allows
attackers to jam wireless networks within a radius of one kilometre using
off-the-shelf equipment.

Affecting various hardware implementations of the IEEE 802.11 wireless
networking standard - including widely used 802.11b devices - the flaw was
found in the collision avoidance routines used to prevent multiple devices
from transmitting at the same moment.

"When under attack, the device behaves as if the channel is always busy,
preventing the transmission of any data over the wireless network," a
security advisory (http://www.auscert.org.au/render.html?it=4091) released
by AusCERT reads.

The weakness allows miscreants to take down networks within five seconds,
according to researchers at Australia's Queensland University of
Technology's Information Security Research Centre (ISRC), which discovered
the vulnerability.

ISRC's leader of network and systems security research, Associate Professor
Mark Looi, whose PhD students, Christian Wullems, Kevin Tham and Jason
Smith discovered the flaw, said any organization that relies heavily on
wireless infrastructure should take the threat seriously.

"Anyone who's relying on the availability of a wireless network should
really consider that their wireless network can be knocked offline at any
time," said Looi. "They need to very seriously evaluate that network and
decide if it's possible to move away from wireless technology."

While previous denial of service attacks against wireless networks have
required specialised hardware and relied on high-power antennas, the new
attack will make knocking a wireless network off the air an option for a
"semi-skilled" attacker using standard hardware.

"An attacker using a low-powered, portable device such as an electronic PDA
and a commonly available wireless networking card may cause significant
disruption to all WLAN traffic within range, in a manner that makes
identification... of the attacker difficult," The AusCERT advisory read.

Because the flaw is in the 802.11 protocol itself, the vulnerability cannot
be mitigated through the use of software or encryption schemes. Replacing
wireless devices with those not affected by the flaw seems the only option,
said Looi.

"Mitigation strategies are few and far between," Looi said "Organisations
could deploy wireless networks that don't use this technology, [but] it
will be a very expensive exercise."

The flaw is only present in devices using a Direct Sequence Spread Spectrum
(DSSS) physical layer, including IEEE 802.11, 802.11b and 802.11g wireless
devices operating at low speed. 802.11a and 802.11g wireless devices
configured to operate at speeds above 20Mbps are not affected by the glitch,

AusCERT senior security analyst Jamie Gillespie does not anticipate the
wide exploitation of the vulnerability.

"For the average corporate user, we're not expecting to see ongoing denial
of service attacks. However, if you have remote equipment that is only
connected through wireless it is possible that the connection could be
disrupted," Gillespie said. "Some critical infrastructure providers may not
deploy wireless... but if any do then they should be looking at mitigation
strategies."

The lack of a "measurable result" during an attack is likely to render the
average attacker bored, Gillespie added.

Unlike flaws discovered in the WEP encryption scheme, the newly-disclosed
vulnerability will not allow an attacker to snoop on network communications.

The ISRC findings will be presented to the Institute of Electrical and
Electronic Engineers (IEEE) Wireless Telecommunication Symposium in
California on Friday.

Archives at: <http://Wireless.Com/Dewayne-Net>
Weblog at: <http://weblog.warpspeed.com>

Wow. I'm underwhelmed. First, we collectively realize that RFC793 really
HAS said 'The RST has to be in the window, not dead on' for 2 decades.

Now somebody's discovered that a jabbering transciever can take out a
CSMA/CA network. And to top it off:

  The model of a shared communications channel is a fundamental
  factor in the effectiveness of an attack on this vulnerability.
  For this reason, it is likely that devices based on the newer IEEE
  802.11a standard will not be affected by this attack where the
  physical layer uses Orthogonal Frequency Division Multiplexing
  (OFDM)

Yes - *THIS* attack doesn't work, you can't take down a 802.11a net
with a hacked PDA that jabbers on channel, you need to use a hacked
802.11a-capable PDA that jabbers on all the subchannels at once. :wink:

Any bets on what will be rediscovered next? Some CERT will realize that
if a DDoS uses RFC1918 source addresses, it will be hard to track down the
misbehaving sources? :wink:

I can think of one application - the next time I'm presenting at a conference where everyone has their heads buried in their laptops, I know what I'll be running on my machine at the podium :wink:

Bill.

I can think of one application - the next time I'm presenting at
a conference where everyone has their heads buried in their
laptops, I know what I'll be running on my machine at the podium
:wink:

what keeps the cows in the pasture is the quality of the grass not
the height of the fence.

randy

Isn't that directly related to the number of cow-chips on the pasture?

On May 14, 2004 Randy Bush spake:

Randy Bush writes on 5/14/2004 7:13 PM:

I can think of one application - the next time I'm presenting at
a conference where everyone has their heads buried in their
laptops, I know what I'll be running on my machine at the podium
:wink:

what keeps the cows in the pasture is the quality of the grass not
the height of the fence.

You can take a horse to water, but you can't make it drink.

yours, etc.

what keeps the cows in the pasture is the quality of the grass
not the height of the fence.

You can take a horse to water, but you can't make it drink.

i am greatly cheered by non-listening competitors.

we have actually watched nanog/ietf/... traffic levels, and one
can clearly tell when there is an interesting talk/speaker; net
use drops non-trivially.

the bottom line is that, if you want people to listen to you,
then have something interesting to say and say it well. and just
being cute doesn't cut it.

randy

## On 2004-05-13 21:43 -0400 Valdis.Kletnieks@vt.edu typed:

Any bets on what will be rediscovered next? Some CERT will realize that
if a DDoS uses RFC1918 source addresses, it will be hard to track down the
misbehaving sources? :wink:

No - then someone would have to re-invent backscatter analysis ... :wink:

Rafi Sadowsky writes on 5/14/2004 11:28 PM:

## On 2004-05-13 21:43 -0400 Valdis.Kletnieks@vt.edu typed:

Any bets on what will be rediscovered next? Some CERT will realize that
if a DDoS uses RFC1918 source addresses, it will be hard to track down the
misbehaving sources? :wink:

No - then someone would have to re-invent backscatter analysis ... :wink:

And someone would then start another thread about BCP 38 on nanog ... funny how several threads turn into a thread about spoofed source address filtering in no time at all :slight_smile:

  srs

Suresh Ramasubramanian wrote:

And someone would then start another thread about BCP 38 on nanog ... funny how several threads turn into a thread about spoofed source address filtering in no time at all :slight_smile:

Let the record reflect the fact that it was not I who did that this
time. I forgot where the grease spot (from the oft-beaten horse) was.

I can think of one application - the next time I'm presenting at a conference where everyone has their heads buried in their

laptops, I know what I'll be running on my machine at the podium :wink:

Bill.

Wayback before laptops, an old hand at IBM stopped in the middle of his presentation and said, "If you guys get done listening
before I get done talking, please let me know so I can leave."

--Michael