Secure model for distributed data structures on distributed hash tables

Abstract

Distributed Hash Tables and unstructured Peer-to-Peer networks present an alternative basis for global software applications. Compared to the Client-Server applications, P2P applications have significant advantages. By eliminating the idea of Servers as the single point of failures and single point of control out of the picture. P2P networks are suitable for many applications like Distributed Online Social Networks the modern society needs today. Distributed Applications built on top of a P2P network are much more resistant to complete system breakdowns, power misuse by authorities (e.g. censorship) and limitless surveillance. In previous work, we presented Distributed Data Structure (DDS) offers a middle-ware for distributed applications. This software has been implemented in a simulation framework for P2P networks called PeerFactSim.KOM. The DDS middle-ware works on top of a Distributed Hash Table (DHT) overlay as a structured P2P network and offers an object-oriented, distributed storage layer. The security of data is an essential topic in such environments. Therefore, without proper security mechanisms, the idea of storing private and sensitive data on unknown network peers become useless, as the stored data can be read or manipulated. In this paper, we present and evaluate a concept of a secure model working completely without trusted nodes for such distributed data structures in peer-to-peer networks. In the evaluation, we show that the time and storage overhead introduced through the security architecture for DDS comes with acceptable proportions for large P2P applications.