Black Hat Europe – day 1

Programme packed with interesting talks.

Though the prestige of

Black Hat Europe

doesn’t compare to that of its American parent conference, and the event certainly doesn’t dominate the debate on

Twitter

in quite the same way, more than 800 security experts descended on Amsterdam this week where, in the RAI Convention Centre, the 14th edition of

Black Hat Europe

is taking place.

The conference opened with a keynote from

Adi Shamir

(perhaps still best known as the ‘S’ in the RSA protocol) on side channel attacks. He started by describing how it is possible for an adversary to extract the private RSA key by measuring the power usage of a computer that uses that key to decrypt data.

Most of Adi’s presentation, however, concentrated on an attack that used a printer/scanner, a laser, and ultimately even a drone to extract data from an air-gapped network after it had been infected with malware. It was a fascinating presentation, even if probably of little practical use for anyone not in the business of writing film scripts.

After the keynote, the conference split into four parallel streams. I stayed in the main room to watch a presentation by

Jose Selvi

on bypassing HSTS.

HSTS (

HTTP Strict Transport Security

) allows a web server that has been contacted over HTTPS to tell the client to force all connections for the next

n

seconds to use HTTPS, for some usually very large value of

n

. This prevents a user who enters the URL manually (or uses a non-HTTPS bookmark) from becoming the victim of a man-in-the-middle attack.

As the value of

n

usually isn’t infinity, Jose demonstrated a weakness in HSTS by performing a man-in-the-middle attack on the NTP protocol, forcing the client’s computer to change its time to a future date. The ‘Delorean’ tool he demonstrated (named after the car in the

Back to the Future

film series) seemed pretty neat and showed that HSTS isn’t a silver bullet for enforcing HTTPS.

Symantec

researcher

Candid Wüest


spoke about

wearable devices at VB2014 and, having missed that presentation, I can see why people were so excited about it. To be filed under the category ‘it would be funny if it wasn’t true’, Candid showed how some of these devices have never been within a mile of a security expert, sending unencrypted and unauthenticated data over the Internet and making many other rather basic mistakes. And while that might perhaps not have surprised many people in the room, Candid demonstrated how he had already been able to track a large number of delegates through their wearables.

Although I have an interest in cryptography, I know very little about quantum cryptography. Hence I was intrigued by the presentation from

BT

‘s Konstantinos Karagiannis, in which he showed how by combining quantum theory and Fourier analysis, in the future one might be able to break RSA keys in real time. Konstantinos also showed a much more positive result: using quantum properties, one can send data (such as encryption keys) over a network with the property that it vanishes as soon as someone looks at it.

Axelle Apvrille

has spoken at and written for

Virus Bulletin

on several occasions in the past (do read her most recent paper on

AdThief

if you haven’t done so already), so I was interested to see her presentation on research she has performed with

Ange Albertini

. In their presentation, they showed how one could hide an

Android

app (such as malware installed by a second, apparently harmless, app) inside a PNG image that would AES-decrypt to the malware.

I was expecting a lot of brute forcing to find a key that would turn the malware into a PNG image, but it turned out that the trick is far more subtle and uses only some basic cryptography, as well as the fact that both PNG images and

Android

packages can contain a large amount of redundant data.

The final talk of the day was also

Android

-related. In a presentation that didn’t shy away from technical details,

Sagi Kedmi

explained that there is a weakness in the pseudo-random number generator used by

Android

, which turns out to be a lot more predictable than it is supposed to be, especially during the first moments after a device is booted up when not enough entropy is available.

As Sagi explained, and later showed in some demonstrations, this has serious consequences and could, for instance, be abused by malware to attack other apps and cause further harm.

The conference continues on Friday with another full day. If Thursday was anything to go by, I expect the presentations to be good!

Posted on 17 October 2014 by

Martijn Grooten