Un-Hackable Computers

Obfuscation is a ten-dollar word for a fairly simple concept. When applied to language, it refers to making a message confusing or difficult to understand through unnecessary wordiness or use of jargon. As it relates to software, obfuscation refers to security measures that prevent tampering or reverse engineering by concealing the purpose or logic of computer code. Programs known as obfuscators are used to transform a readable program into obfuscated code, and only by knowing the key can the original program be unlocked.

About two years ago, two independent groups of researchers came up with an approach that uses complicated mathematical operations to keep a source code secret. Called “indistinguishability obfuscation” (IO), it was touted to be unbreakable. But IO had two major problems. First, performing all of the mathematical operations needed to achieve true indistinguishability would have taken years…in fact, hundreds of years. Second, further research concluded that it was not as secure as originally thought.

In recent months, however, much headway has been made toward making IO a practical reality. By breaking large, open-ended computations into smaller, discrete steps and tying them together at the end, an entire program can be protected in a shorter period of time. In addition, a technique from lattice-based cryptography known as “learning with errors” (LWE) is being evaluated as a way to make IO truly non-hackable. At the pace things are progressing, the team believes that near-perfect computer security may be possible in as little as 10-15 years.