Solid Basis for Symmetric Cryptography

Opis projektu

Solidne fundamenty kryptografii symetrycznej

W ciągu ostatnich dwóch dekad kryptografia symetryczna znacznie się rozwinęła za sprawą prac nad jej standaryzacją. Bezpieczeństwo kryptografii jest jednak uzależnione od kryptoanalizy, co wymaga ciągłej kontroli w celu poprawy zaufania do funkcji kryptograficznych. Finansowany przez ERBN projekt SoBaSyC zakłada ustanowienie solidnych fundamentów dla kryptografii symetrycznej poprzez konsolidację zgromadzonych przez lata wniosków dotyczących różnych metod ataku, udoskonalenie ich poprzez podejście algorytmiczne oraz optymalizację ich zastosowania. Inicjatywa ma doprowadzić do opracowania zestawu narzędzi zawierającego nowo zaproponowane i zoptymalizowane algorytmy, zdolne do przeprowadzania najskuteczniejszych znanych ataków na daną konstrukcję. Dodatkowo, z tej metodologii algorytmicznej wyprowadzone zostaną teoretyczne ograniczenia i właściwości, co pozwoli poprawić jej pewność i efektywność przy jednoczesnej oszczędności czasu.

Cel

Symmetric cryptography, essential for enabling secure communications, has benefited from an explosion of new results in the last two decades, in big part due to several standardization efforts: many public competitions have been launched since 1997, where the community proposes cryptographic constructions and simultaneously evaluates their security and performance. The security of symmetric cryptography is based on cryptanalysis: we only gain confidence in a symmetric cryptographic function through extensive and continuous scrutiny.
However, the current context has not allowed the community to digest all the new findings, as can be seen from several recurrent issues. The two main ones are:
1) primitives proposed at top-tier venues often get broken by slight modifications of already known techniques;
2) published cryptanalysis at top conferences sometimes include mistakes or are suboptimal. They are also often re-invented and re-named.

The main challenge of SoBaSyC is to establish solid bases for symmetric cryptography. Using cryptanalysis as the starting point, my aim is to unify the knowledge obtained through the years on the different families of attacks, to transform it with an algorithmic approach and to endow it with optimizations. The final result will be a toolbox congregating all our newly proposed optimized algorithms, that will provide the best known attacks on a given construction, through an easy application. Next, I plan to derive from this algorithmic approach some theoretical bounds, as well as some properties that I will include in the security proofs of symmetric constructions, providing more meaningful and realistic security arguments.

This would allow, for the first time, to ensure that any newly proposed primitive or construction is already resistant to all known attacks, and will considerably increase the confidence on these functions. It will also save a considerable amount of time and allow the field to advance, at last, on solid ground.