Skip to main content

DOTANDO DE CIBERSEGURIDAD SOSTENIBLE A LA INTERNET DE LAS COSAS - SECURING/NET

PID2021-125962OB-C33

•
Funding agency name Agencia Estatal de Investigación
Funding agency acronym AEI
Program Programa Estatal para Impulsar la Investigación Científico-Técnica y su Transferencia
Subprogram Subprograma Estatal de Generación de Conocimiento
Call Proyectos de I+D+i (Generación de Conocimiento y Retos Investigación)
Call year 2021
State Plan 2021-2023 Plan Estatal de Investigación Científica y Técnica y de Innovación
Beneficiary centre UNIVERSIDAD AUTONOMA DE BARCELONA

Resultados relacionados

Resultados totales (Incluyendo duplicados): 3
Encontrada(s) 1 página(s)

O2, repositorio institucional de la UOC
Publication Contribución de congreso (ConferenceOutput)

Privacidad en Smart Grid mediante Smart Contracts y agregación de datos

  • Ferré-Queralt, Joan
  • Anglés-Tafalla, Carles
  • Salas, Julián
  • Navarro-Arribas, Guillermo
  • Viejo, Alexandre
  • Megias, David
  • Castellà-Roca, Jordi
En este trabajo presentamos un sistema para la gesti´on de redes de distribución de energía inteligentes (Smart Grids) donde se preserva la seguridad y privacidad de los usuarios. La propuesta utiliza tecnología blockchain y Smart Contracts como componentes centrales, teniendo en cuenta la privacidad de los usuarios, la disponibilidad y la seguridad del sistema. Al mismo tiempo, ofrece medidas para identificar y tomar acciones punitivas contra usuarios maliciosos. El sistema proporciona, finalmente, informaci´on de consumo y producción de energía para su publicación o uso, donde se preserva el anonimato de los usuarios con el principio de k-anonimidad.
O2, repositorio institucional de la UOC
Publication Contribución de congreso (ConferenceOutput)

Unmasking the Illusion: The Shortcomings of “Zero-Knowledge” Rollups in Achieving Privacy

  • Torralba Agell, Adrià
  • Keshavarzkalhori, Ghazaleh
  • Pérez-Solà, Cristina
  • Megias, David
  • Herrera-Joancomartí, Jordi
The rise of Layer 2 (L2) solutions, including Payment Channel Networks, sidechains, and rollups, has aimed to tackle the scalability challenges of Layer 1 (L1) blockchains like Bitcoin and Ethereum. Among these, Zero-Knowledge Rollups (ZK-Rollups) have emerged as a compelling solution by utilizing ZK-SNARKs to bundle multiple transactions, thereby enhancing throughput and reducing costs. However, despite their technical sophistication, ZK-Rollups do not inherently provide transaction privacy, a common misconception given the “Zero-Knowledge” nomenclature. This paper explores the privacy limitations of ZKRollups, emphasizing the need for privacy-preserving features that align with the expectations set by their name. We also review the strategies being eveloped by various projects to address these limitations. Furthermore, we propose the community begin adapting other names for the technology, such as “Verifiable Rollup” (verRollup), “Incrementally Verifiable Computation Rollup” (ivcRollup), or “Succinct Rollup” (sucRollup) that better represent the current capabilities of rollups. This work contributes to the ongoing discussion on achieving a balance between efficiency, scalability, and privacy in blockchain technologies.
O2, repositorio institucional de la UOC
Publication Contribución de congreso (ConferenceOutput)

Unveiling Privacy Risks: Property Inference Attacks in Federated Learning Environments

  • Serebriakova, Antonina
  • Pineda Burgueño, Gorka
  • Garcia-Font, Victor
  • Pérez-Solà, Cristina
Federated Learning (FL) advances privacypreserving machine learning by decentralizing model training to individual devices, ensuring data remains localized. However, FL is not immune to privacy threats. This paper investigates the specific risk of property inference attacks, where an adversary infers whether a certain property is present in the training data, despite this property not being the focus of the global model. Through experiments with TensorFlow Federated, we replicate and extend previous findings on property inference attacks, validating their reproducibility, credibility and robustness. Our analysis reveals two key insights: property inference attacks are highly effective during the initial training rounds due to significant early updates, and the rarity of the property in the dataset enhances attack effectiveness due to greater weight contrast. Our results demonstrate that even without adversarial modifications, significant privacy risks exist in FL systems, highlighting the need for enhanced security measures to protect sensitive information in FL environments.