Working to confuse
Programme obfuscators are algorithms that process input data so that the result for intended viewers is different to that for non-intended viewers — perhaps even making it incomprehensible. Viewing an obfuscated programme should not offer insight or ability knowledge beyond what is needed to simply run the programme on inputs of choice. Although attempts to obfuscate programmes often fail, recent work in this area has shown that it is possible to apply provably secure obfuscators for certain simple classes of programmes. The 'New directions in cryptographic program obfuscation' (Crypto obfuscation) project is working to enhance understanding of the concept of programme obfuscation and its applicability to cryptography. The EU-funded project intends to broaden the boundaries of what is actually obfuscatable by proposing new methods for obfuscating richer classes of programme. Team members believe they can formulate and achieve stronger notions of obfuscation than those currently being considered, and find ways to work around known impossibility results for general-purpose obfuscators. Another objective is to uncover new applications of programme obfuscation in other areas of cryptography. Project achievements aim to enhance the worth of applying obfuscation to real life scenarios and to offer obfuscation algorithms that can be embedded in real life systems. In work to date, project partners have come up with a new notion of obfuscation, grey box security, and have put forward new obfuscation algorithms for advancing work at the algebraic level. Studies have been performed on connections between multi-bit point obfuscation as well as strong forms of encryption that are secure against information being leaked. In other achievements, the Crypto obfuscation project has introduced the strong notion of non-malleable obfuscation, and proposed new obfuscation algorithms that can bear up to related requirements. To date, project partners have published five papers, with others being prepared for submission.