I was thinking about writing something on encryption, specifically the Chinese Remainder theorem, which is based on ancient Chinese/Hindu mathematics, and contributes to the foundation of modern encryption. So I got out my old discrete mathematics textbooks and started reviewing things.

So it turns out that I used to be a lot smarter than I am now. Solving systems of linear congruencies, fiddling with probabilistic primality tests, contemplating moduli – this makes my head hurt.

Sun Tzu, the third century mathematician (not the same as the Art of War guy) offered the following puzzle:

There are certain things whose number is unknown. When divided by 3, the reminder is 2; when divided 5, the remainder is 3; and when divided by 7, the remainder is 2. What will be the number of things?

Doesn’t that make your head hurt?

Luckily, we don’t have to figure all this out from scratch - here at Mozy we use the open_ssl implementation of Blowfish to encrypt everything. And since it’s open source, it’s subject to the scrutiny of the ever-vigilant tin-foil hat crowd.

We’re a little unique here in that we give our users the option to provide their own private key (a passphrase, a picture, a music file – whatever) which is hashed to generate a 448-bit key for encryption. This key is stored on their PC, and we don’t have access to it – it’s stored on the PC and is never transferred to our servers – which means that we can confidently protect data from both kid sisters and government agencies.

To help explain how strong the encryption is, let’s consider if you were to brute force an attack on the data encrypted on the Mozy servers. You’d have to consider 2⁴⁴⁸ different keys, which is a lot of keys (the numbers of atoms in the universe is estimated to be somewhere around 2⁷⁷ or so.)

But let’s say the DOE Lawrence Livermore National Laboratory let you borrow BlueGene/L, which is the most powerful computer in the world (as of this writing) and let’s assume that the Linpack Rmax Gflops metric is a proxy for key generation – where a key can be generated and tested in about the same amount of time it takes to perform 1,000 floating point operations. So that’s 280,600,000,000 keys per second.

This means that it would take BlueGene/L about 30 years to guarantee that it would crack the encryption that Mozy utilizes. Now, this doesn’t mean that Mozy is all that when it comes to encryption – it’s not like we invented this algorithm – we just selected it based on its strength and speed.

So that’s BlueGene/L – but if you used your own PC, then it’d take closer to 3,026,328,640 years or so. Or if you got a bunch of people to cluster your computers together – say 10,000 of your closest friends, it’d still take about three hundred thousand years.

But, since we’re completely paranoid here at Mozy, we not only used 448 bit encryption, but we also go the extra mile and carefully secure our data disks with aluminum foil, which approximates a Faraday cage. This protects them from electromagnetic radiation as well as potential telekinetic security breaches.

Below is a quick demonstration of the process – but note that we have filed a patent on this process, and will vigorously defend any infringement on our intellectual property.

Step 1: Choose your foil.

This is a tough one. On the one hand, you’ve got Reynolds – which is an old standby. On the other hand, perhaps it’s safer to use a more generic brand, like from Target. Reynolds is made in Virginia, which is home to all sorts of government agencies, and who knows how they’ve meddled with the manufacturing process of this aluminum foil. But the Target brand is manufactured in Minneapolis, which was the home of the Cray supercomputer, and everyone knows they were in cahoots with the government, specifically the NSA. Also, that dark blue packaging color is strangely similar between the two companies…

So, we’ll take our chances with Reynolds.

Step 2: Envelop the disk in the foil.

This is relatively straight forward. Be careful not to tear the foil on the sharp edges of the disk encasement. We recommend the duller side of the foil to be on the outside as we have found it to have better diffusive properties than the shiny side.

Small openings can be made for cable attachments.

The answer to the puzzle is 23.