The principle of open design. This is the best known of
the design principles, and it's also
the least understood. If you look at the next slide, open design simply says
security should not depend upon secrecy of design or
implementation. That's it. Programs the security
does depend on, this are often called
security by obscurity. It's very bad because that means if the secret is
perceived as broken, you're wide-open, you're
completely vulnerable. Now, this applies to
design or implementation. This does not mean you should
publish cryptographic keys. So keeping small pieces of
information secret is fine, like a password or
a cryptographic key. It's when you hide the design or the implementation that uses those keys that things are bad. Now, one other thing, it's popularly believed
or widely believed that this means source code should be made public or things
should be made public. No. All this is saying is that your security should not depend solely on secrecy of
design or implementation. It's perfectly fair
to have that as one of a number of
characteristics. But you need to be sure
that if someone does discover the design or
the implementation, that you're so secure. That's all it means. There's a wonderful story
illustrating this, involving the DVD Copyright
Control Association. If you ever looked at a DVD, you know it's encrypted. It's protected by what's
called the Content Scrambling System or CSS. Now, this was secret. Part of the reason
for being secret was because they didn't want anyone
to know how it was done. So people couldn't
break it and so forth. Here's how it works. First of all, the disc has
the layout that's shown there, K_a hash, K_d, and so forth. Now, when it starts up, the first thing the player does is it checks the authentication
key to read that in. Now, embedded in the player key, rather embedded in the player, is its own key. This is unique by manufacturer. So then if you look at
the third block on, you'll see that the K_d, which is the disc key, is enciphered using
a large number of player keys from the third
to the one at the bottom, next to last one. So what the system does is it goes to the slot that
it's been assigned, uses the player key to decipher
the K_d, the disc key. What it does then is hash that and compare
it to the hash. In some versions, what
they'll simply do is go down all of the
encrypted data keys, find the associated disc key, hash that and compare
it to the hash. When there's a match,
they've got the right key. You then use the K_d on
the last, the very bottom, you take that, and you
use that to decrypt K_t, which is the title key. Then you can play your DVD. Well, the algorithm to
do this is complex. What happened was in 1999, a group in Norway
derived an algorithm that was completely
compatible with CSS. They did this because
up until then, you could not play DVDs on Linux or encrypt
the DVDs on Linux. By doing this, they were
able to come up with an algorithm that would allow
you to play DVDs on Linux. The software started
spreading rapidly, and the DVD Copyright
Control Association sued in the US courts to prevent
further distribution. Basically, they wanted it not
distributed within the US. They were also suing in
Norway for something similar. In order to do this,
they had to show that the program of the code that the Norwegian people developed, the Norwegians developed
was in fact the CSS. So what they did was they filed an affidavit containing
the source code to the CSS and said, "Look, these are the same." But their source code, the DVD Copyright Control
Association source code was a trade secret. They filed it with the court
in Santa Clara, California. That court at the time
was experimenting with putting pleadings and
information on the web. So it went right out on
the web like everything else. One day later, the lawyers
realize that they hadn't asked it be sealed as they
should for trade secrets. So they went to the court
and asked it be sealed. It was sealed. So no one except the lawyers could read it. The affidavit was
pulled off the web. But it had been up
there for one day. It had been downloaded
roughly 20,000 times too. So this was needless to say, spread among a lot of people. The cryptographers
had a field day because the cipher
is incredibly weak. So as a result, there were a number
of breaks to it. Now you can find the code. It's fairly widely
distributed on the Internet. In fact, you can even download programs that will
play encrypted DVDs. That's a good example where the Copyright Control
Association apparently relied on secrecy of the crypto algorithm in order to protect
their security. Good idea, very tempting to
do, very common mistake. The problem is
humans are fallible. The lawyer slipped up. As a result,
the secrecy was gone. You also have to be
careful about aggregation, pulling different pieces of information together for this. As an example,
during the Gulf War, the Associated Press
was looking at various US defense
contractors sites, and suddenly realized that there was a lot of information on those sites that
they could use to figure out where
troops were located. So they went out and looked at a number of websites,
and from that, we're able to construct
the location of many US military bases in Iraq. That information was
very, very secret. So the first thing
they did was they went to the defense
contractors and said, " Do you know we can do this." The contractors were horrified, and they pulled the information
right off the Internet. Then the AP published its story. My point here is that this information may become available in ways
you don't expect. So don't rely on secrecy. It can be one step, but it can't be the only step. On that note, let's go onto the next video for this lesson.