Constantinos Patsakis, Agusti Solanas

IACR Cryptol. ePrint Arch. 2013, 489

Description: The increasing demand for on-line collaborative applications has sparked the interest for multicast services, which in many cases have to guarantee properties such as authentication or confidentiality within groups of users. To do so, cryptographic protocols are generally used and the cryptographic keys, in which they rely, have to be managed (eg created, updated, distributed). The procedures to perform these operations are determined by the so-called Group Key Management Schemes. Many schemes have been proposed and some of them have been proven to be vulnerable. This is the case of the Piao et al. scheme, whose scalability/efficiency is very good but it is vulnerable to many attacks because its security is based on a weak mathematical problem, so it can be broken in polynomial time. Inspired by the concepts proposed in the Piao et al. scheme we have redesigned the protocol and we have founded it on a hard mathematical problem and tweaked some of the procedures. This way, we propose a new scheme that is efficient, collusion free, and provides backward and forward secrecy.

Cesar Mauri, Agusti Solanas, Toni Granollers, Narcís Parés, Joan Bagés, Mabel García

IV Jornadas Iberoamericanas de Tecnologías de Apoyo a la Discapacidad, 35-43

Description: El presente artículo describe las estrategias y tecnologías utilizadas para proveer a personas con parálisis cerebral en grados moderados y severos de unas actividades lúdicas y creativas diseñadas a la medida de sus potencialidades. Estas actividades se basan en sistemas interactivos de visión artificial y generación de gráficos y sonidos en tiempo real. El bienestar conseguido con estas actividades parte de conseguir un importante grado de autonomía en su realización gracias a las tecnologías utilizadas, las cuales no requieren la mediación de un cuidador o educador para ayudar a los usuarios a realizarlas.

Josep Domingo-Ferrer, Francesc Sebé, Agusti Solanas

Proceedings of Joint UNECE/Eurostat work session on statistical data confidentiality,

Description: p-Sensitive k-anonymity is a sophistication of k-anonymity requiring that there be at least p different values for each confidential attribute within the records sharing a combination of key attributes. Like for k-anonymity, the computational approach originally proposed to achieve this property is based on generalization and suppression; this has several data utility problems, such as turning numerical key attributes into categorical, injecting new categories, injecting missing data, etc. We present and evaluate two heuristics for p-sensitive k-anonymity which, being based on microaggregation, overcome most of such drawbacks, while offering a smooth information loss increase as p and k grow.

Page 9 of 43

Av. Països Catalans, 26, 43470 Tarragona, Catalonia, Spain

Av. Països Catalans, 26, 43470 Tarragona, Catalonia, Spain
Latest from our Blog