Secure and decentralized usercentric network architecture

Jernej Kos (2016) Secure and decentralized usercentric network architecture. PhD thesis.

Preview

PDF Download (2636Kb)

Abstract

We live in an increasingly interconnected world where more and more interactions depend on digital communication services that include millions of devices connected to a common network such as the Internet. However, at the same time the model of communication between users or their applications is unconsciously changing – the Internet was designed as a decentralized network resilient to failures and attacks, a network in which users can directly exchange information, but more and more communication services are becoming dependent upon large service providers. On one hand, this is understandable, since centralization brings advantages in managing server infrastructure, while businesses that manage these services easily attain higher earnings. However, for the users of these services this often means invisible loss of control over their personal data, which is now stored in data centers in different locations scattered around the world, in countries with different privacy protection laws and in particular out of the reach of users. By using such centralized databases that include personal information of a large number of users, such collections become tempting targets for attackers and various intelligence services are already exploiting this fact. In this thesis, we therefore start with the construction of an alternative network architecture that could serve as a new way of communication between users, which emphasizes control over the network infrastructure and the data that is transmitted over it. To this end, we develop a novel routing protocol, which allows users to form an independent network and is resistant to attacks on the routing infrastructure. Networks in the real world often contain large numbers of users, which means that the protocol must be scalable in order to be useful. We prove that the developed protocol, with high probability, places an upper bound of

Actions (login required)