Paper 2019/453

A New Approach to Modelling Centralised Reputation Systems

Lydia Garms and Elizabeth A. Quaglia

Abstract

A reputation system assigns a user or item a reputation value which can be used to evaluate trustworthiness. Blömer, Juhnke and Kolb in 2015, and Kaafarani, Katsumata and Solomon in 2018, gave formal models for \mathit{centralised} reputation systems, which rely on a central server and are widely used by service providers such as AirBnB, Uber and Amazon. In these models, reputation values are given to items, instead of users. We advocate a need for shift in how reputation systems are modelled, whereby reputation values are given to users, instead of items, and each user has unlinkable items that other users can give feedback on, contributing to their reputation value. This setting is not captured by the previous models, and we argue it captures more realistically the functionality and security requirements of a reputation system. We provide definitions for this new model, and give a construction from standard primitives, proving it satisfies these security requirements. We show that there is a low efficiency cost for this new functionality.

Metadata
Available format(s)
PDF
Category
Cryptographic protocols
Publication info
Published elsewhere. Major revision. 11th International Conference on Cryptology, AFRICACRYPT 2019
Keywords
group signaturesdirect anonymous attestationreputation systems
Contact author(s)
Lydia Garms 2015 @ rhul ac uk
History
2019-05-08: received
Short URL
https://ia.cr/2019/453
License
Creative Commons Attribution
CC BY

BibTeX

@misc{cryptoeprint:2019/453,
      author = {Lydia Garms and Elizabeth A.  Quaglia},
      title = {A New Approach to Modelling Centralised Reputation Systems},
      howpublished = {Cryptology ePrint Archive, Paper 2019/453},
      year = {2019},
      note = {\url{https://eprint.iacr.org/2019/453}},
      url = {https://eprint.iacr.org/2019/453}
}
Note: In order to protect the privacy of readers, eprint.iacr.org does not use cookies or embedded third party content.