HyperNet: A conditional k-anonymous and censorship resistant decentralized hypermedia architecture

Abstract

Nowadays, the vast majority of Internet services used to distribute hypermedia content follow a centralized model, which is highly dependent on servers and raises several quality and security concerns. Among other issues, this centralized model creates single points of failure, requires trust on providers to avoid censorship and personal data misuse, and results in a scenario where digital content tends to disappear or be inaccessible over time, for example, when a content creator stops maintaining a site or when the content is moved to another location. To improve this, it is necessary to replicate data and follow more distributed models. Nevertheless, current platforms to distribute content in this way, either do not offer an effective mechanism to maintain the privacy of their users or they offer full-anonymity, which contributes to the dissemination of content that goes beyond the law and moral standards of many users.

This paper proposes a novel distributed architecture that enables hypermedia resource distribution ensuring censorship resistance and conditional k-anonymity. In the proposed system, users form groups to share hypermedia content where the anonymity of the publisher is preserved only if the publication follows a set of rules defined by the group. To this end, the proposed system uses threshold discernible ring signatures to enable conditional k-anonymity, the Ethereum blockchain platform to manage groups and user identities, and the InterPlanetary File System to store and share hypermedia resources in a distributed way. This document provides the design for the proposed architecture and protocols, it evaluates system risks and its security properties, and it discusses the proposal in general terms.