.
Bitland.Net Security Notes            Comments? email jwilkins-at-bitland*net
More information on the author at Jonathan Wilkins's home page
RSS feed available at http://www.bitland.net/index.rss               Add to Google
Archives: 2007, 2006, 2005, 2004, 2003, 2002, 2001, 2000
Glitch Attacks and Amateur Cryptographers  |  (2007/05/11 12:05)

Nate's blog (rdist) talks about Glitch Attacks and links to some good papers on the topic. Basically, this is probably the most interesting general purpose cryptographic attack technique to come about in the past few years. By introducing fluctuations to the power supply or clock you can make the CPU execute a number of wrong instructions. If you can do this, you can recover a DES key with between 1 and 10 faulty ciphertexts. You can factor a RSA key with one.

The most fascinating thing is that this technique seems to have originated with the pay-tv hacking community, not from academics or conventional crypto researchers.
+digg  |  +del.icio.us   |    [Crypto ]   |   Permanent link
Fourmilab Releases Ent - Randomness Testing Utility  |  (2007/01/20 17:00)

Ent is available here and looks like a good alternative to Diehard or the NIST STS suite.
+digg  |  +del.icio.us   |    [Crypto ]   |   Permanent link
Cache Timing Attacks on AES  |  (2005/06/03 16:45)

Dan Bernstein has recently released a paper on Cache-timing attacks on AES. His attack is able to perform complete AES key recovery based on known plaintext timings across a network.

This is pretty huge and is fundamental to the AES algorithm, not any particular implementation.

Tom Ptacek has an astoundingly funny edit of the resulting debate on sci.crypt.
+digg  |  +del.icio.us   |    [Crypto ]   |   Permanent link
Bernstein's paper on Brute Force Attacks in Cryptography  |  (2005/06/03 15:00)

I love reading DJB's papers. They're always quite educational and he's not above handing out a well deserved smackdown. Take this quote from his Understanding Brute Force paper. 
"Why did the author of this paper characterize this serial attack as successful
cryptanalysis?  It's simply not true that the attack is 'faster than exhaustive
search' - unless you assume that the attacker is forcing himself to use a 
serial computer, ie, that the attacker is an idiot"
+digg  |  +del.icio.us   |    [Crypto ]   |   Permanent link
$20 for an SSL Cert  |  (2005/05/24 12:15)

GoDaddy is selling $20 SSL certificates for a limited time. Verisign is still making an obscene amount of money flogging them for $349/year. All you're paying for is for them to sign a few bits so that browsers will be able to see that a third party has vouched for your site.
Over the past few years, other companies have emerged to sell certs for less than Verisign. Thawte sells them for $149. GeoTrust sells them for $99. A little while ago FreeSSL.com started offering them for $49. The only real difference between them is whether their root certificate is in all the browsers your site is likely to see. If you have a lot of users who are still running Netscape 1.0, then you'll need to pay Verisign to make sure that they don't get any error messages. Otherwise, almost everyone else is fine with the cheap certs you can get elsewhere.
+digg  |  +del.icio.us   |    [Crypto ]   |   Permanent link
CGD for OpenBSD  |  (2003/11/13 18:20)

As mentioned before, I'd been planning on testing out CGD for OpenBSD. I finally got my lazy ass around to it and set it up on OpenBSD 3.4. There were a couple of minor Makefile tweaks necessary, but I've got it up and running with a 50GB partition. So far it's working well. If another week or two of testing goes well, I'm going to redo all of my Linux/Loop-AES installs and replace them with OpenBSD/CGD.
+digg  |  +del.icio.us   |    [Crypto ]   |   Permanent link
Cryptographic Filesystems  |  (2003/05/03 10:50)

Linux's LoopAES is currently my favorite disk crypto system. I've been using it on my main fileserver for about 6 months with no problems. I run Mandrake and it comes with support for LoopAES out of the box.
When I was initially investigating disk crypto, I found a Survey of disk crypto systems which, while a little old, was quite helpful.
I used to use CFS on FreeBSD but had some trouble with corruption. Also, since I had a large RAID array, fscking after a reboot sometimes took hours.
Other people claim to have had better luck with TCFS
OpenBSD's disk crypto is fairly primitive at the moment. The FAQ recommends that you use vnconfig, but that's limited to 2GB, which is really too small.
However, when I was at CanSecWest this year, Theo told me to ask Ted Unangst (one of the OpenBSD guys) about disk crypto and he pointed me at his page on Cryptographic Disk Device for OpenBSD which is supposed to support large disks. I haven't tried it yet but figured I'd link it in case someone else finds it useful.
UPDATE:
CGD for OpenBSD is based on CGD for NetBSD which was written by Roland Dowdeswell. You might want to read his USENIX paper on CGD
+digg  |  +del.icio.us   |    [Crypto ]   |   Permanent link
Quick HOWTO on installing CFS on FreeBSD  |  (2001/01/28 12:00)

CFS was written by Matt Blaze and is described in full Here For a quick installation guide under FreeBSD, look Here

+digg  |  +del.icio.us   |    [Crypto ]   |   Permanent link
RSS feed available at http://www.bitland.net/index.rss