Privacy: Vitalik Buterin relaunches the challenge of cryptographic obfuscation

Confidentiality is becoming an increasingly important area of ​​research for modern cryptography now. On Friday, Vitalik Buterin published the third part of his series dedicated to cryptographic obfuscation. This time, he studies “local mixing”, a method which abandons networks and elliptic curves. Its approach is rather inspired by hash functions and symmetric cryptography. This track hides the internal logic of a program while retaining its operation, with implications for privacy on blockchain.

Another cryptographic route to obfuscation

In his first two texts, Vitalik Buterin, founder of Ethereum, studied constructions based on networks. There first part was published on June 29. There seconddedicated to the iO diamond, was released on July 28. With this part, he changes direction. Local mixing does not use elliptic curves, prime factorization, or networks in its design.

According to him, this method is close to symmetric cryptography. It serves as the basis for encryption and hash functions. The principle consists of working on a circuit made up of logic gates like XOR, AND and NOT. The process preserves the output, while erasing traces of the internal logic.

The circuit goes through several stages before obfuscation. Vitalik Buterin describes, in his last articleincluding reversibility, hardening, gadgetization and mixing. During this phase, unnecessary doors complicate its analysis. The layout changes, while some blocks are replaced by others producing the same calculation.

Vitalik Buterin focuses on reversibility and scrambling

Reversibility plays a central role in the functioning presented. A reversible gate can be rewritten with another set of reversible gates with identical behavior. It facilitates these transformations. Conversely, an AND or OR gate offers fewer possibilities for applying this type of rewriting.

Mixture alone is not enough to hide the structure of the program. So the other steps have to do a lot of the work. Hardening and gadgetization reinforce the transformation before its final obscuration. The objective remains to preserve the calculation, while making its internal logic difficult to find.

This method seeks to go beyond a simple visual modification of the code. The circuit becomes progressively more complex to examine. Confidentiality is then based on the difficulty of understanding its organization despite the conservation of its outputs. This logic places circuit transformation at the heart of the approach.

A still risky avenue for obfuscation

Vitalik Buterin describe local mixing as a “risky and daring bet”. This reservation comes in particular from its link with numerous failed attempts at white-box cryptography. The authors believe that additional efforts could strengthen this approach. They also accept a higher cost for a more robust construction.

Artificial intelligence appears in this reflection. It could condense three decades of hash function development into just a few years. This possibility distinguishes local mixing from network-based approaches. In these approaches, the trade-off relies on security assumptions.

For him, obfuscation thus represents a particularly difficult frontier in cryptography. He called her “cryptography’s final boss” in his June publication. Some very rigorous constructions also display execution times longer than the lifespan of the universe. So the question is about safety and feasibility.

The interest in this research comes mainly from its applications. Obfuscation can transform a program into an encrypted version, while retaining classic inputs and the same outputs. Associated with a blockchain, it could contribute to private votes resistant to collusion. It could then reduce the need to trust an M-of-N committee to guarantee the process.

What follows will above all allow us to assess whether local mixing can exceed its current limits. Vitalik Buterin maintains a cautious approach to a technique still associated with several difficulties. Advances in cryptography and AI, however, could change the capabilities available. Viability will therefore depend on its safety, cost and effectiveness.