On codes for traceability schemes: constructions and bounds
- Autores: José Moreira Sánchez
- Directores de la Tesis: Marcel Fernández Muñoz (dir. tes.)
- Lectura: En la Universitat Politècnica de Catalunya (UPC) ( España ) en 2013
- Idioma: inglés
- Tribunal Calificador de la Tesis: Josep Domingo i Ferrer (presid.), Josep Cotrina Navau (secret.), Maria Bras Amorós (voc.)
- Materias:
- Enlaces
- Tesis en acceso abierto en: TDX
- Resumen
A traceability or fingerprinting scheme is a cryptographic scheme that facilitates the identification of the source of leaked information. In a fingerprinting setting, a distributor delivers copies of a given content to a set of authorized users. If there are dishonest members (traitors) among them, the distributor can deter plain redistribution of the content by delivering a personalized, i.e., marked, copy to each user. The set of all user marks is known as a fingerprinting code. There is, however, another threat. If several traitors collude to create a copy that is a combination of theirs, then the pirated copy generated will contain a corrupted mark, which may obstruct the identification of traitors. This dissertation is about the study and analysis of codes for their use in traceability and fingerprinting schemes, under the presence of collusion attacks. Moreover, another of the main concerns in the present work will be the design of identification algorithms that run efficiently, i.e., in polynomial time in the code length. In Chapters 1 and 2, we introduce the topic and the notation used. We also discuss some properties that characterize fingerprinting codes known under the names of separating, traceability (TA), and identifiable parent property (IPP), which will be subject of research in the present work. Chapter 3 is devoted to the study of the Kötter-Vardy algorithm to solve a variety of problems that appear in fingerprinting schemes. The concern of the chapter is restricted to schemes based on Reed-Solomon codes. By using the Kötter-Vardy algorithm as the core part of the identification processes, three different settings are approached: identification in TA codes, identification in IPP codes and identification in binary concatenated fingerprinting codes. It is also discussed how by a careful setting of a reliability matrix, i.e., the channel information, all possibly identifiable traitors can be found. In Chapter 4, we introduce a relaxed version of separating codes. Relaxing the separating property lead us to two different notions, namely, almost separating and almost secure frameproof codes. From one of the main results it is seen that the lower bounds on the asymptotical rate for almost separating and almost secure frameproof codes are greater than the currently known lower bounds for ordinary separating codes. Moreover, we also discuss how these new relaxed versions of separating codes can be used to show the existence of families of fingerprinting codes of small error, equipped with polynomial-time identification algorithms. In Chapter 5, we present explicit constructions of almost secure frameproof codes based on weakly biased arrays. We show how such arrays provide us with a natural framework to construct these codes. Putting the results obtained in this chapter together with the results from Chapter 4, shows that there exist explicit constructions of fingerprinting codes based on almost secure frameproof codes with positive rate, small error and polynomial-time identification complexity. We remark that showing the existence of such explicit constructions was one of the main objectives of the present work. Finally, in Chapter 6, we study the relationship between the separating and traceability properties of Reed-Solomon codes. It is a well-known result that a TA code is an IPP code, and that an IPP code is a separating code. The converse of these implications is in general false. However, it has been conjectured for some time that for Reed-Solomon codes all three properties are equivalent. Giving an answer to this conjecture has importance in the field of fingerprinting, because a proper characterization of these properties is directly related to an upper bound on the code rate i.e., the maximum users that a fingerprinting scheme can allocate. In this chapter we investigate the equivalence between these properties, and provide a positive answer for a large number of families of Reed-Solomon codes.
