Masterarbeit
Migrating EigenTrust to the Redecentralized Web
Research Area
Students
Anutam Majumder
Advisers
Description
The Semantic Web technology is a re-imagined web, which at its core is decentralized in nature. Naturally a Redecentralized Web brings up the question of establishing and calculating trust relationships among participating nodes. It also brings up the question of evaluating trust score of unknown nodes before their data is to be trusted. The challenge that lies at large is how to compute and evaluate trust scores of participating nodes in a completely decentralized network. EigenTrust, an already existing algorithm for evaluating Trust Scores in peer-to-peer networks is being explored in our thesis to establish and evaluate trust scores in the context of Redecentralized Web.
The current web includes unannotated user data in abundance, mainly controlled by large tech companies. The Redecentralized Web is an attempt to decentralize the web with humans and machines envisioned to be acting in co-operation with each other. Such a decentralized architecture will require intelligent web agents to make use of data and execute operations on their own when requested. It is not a separate entity, but an extension to the current Web. Naturally, these agents need a way to verify if data from other parties can be trusted. In context to the Semantic Web, data at its most basic form can be viewed as an RDF graph or a triple, with each triple having its own Subject, Object and Predicate.
The intention of this master’s thesis is to bring EigenTrust algorithm to the Redecentralized Web and assess its performance based upon several scenarios. We propose to use RDF triple to let each node store its own trust score based on EigenTrust. While calculating the global trust score for each node, we can make use of SPARQL query language to verify the trust score of each node in the network and wait for the network to converge. The network is hereby the peers a web application communicated with. We propose to make use of aTLAS (a Trust Laboratory of Multi-Agent Systems) testbed which is an open-sourced web-based test wizard to setup the environment and conduct the trust score computations based on different scenarios. We are also interested in verifying how EigenTrust algorithm scales for Redecentralized Web according to various factors such as Direct Interactions, Indirect Interactions, Feedback, Recency and Network Structure. Verifying how the EigenTrust algorithm scale can be of importance as EigenTrust has proved that it can scale very well in large peer-to-peer networks. It will be interesting to find out how such an algorithm perform in the context of the Redecentralized Web and can contribute towards establishing trust, given the trustless nature of the Web.
The objective of this master’s thesis is to find an approach or a combination of approaches to solve the mentioned problems within the context of migrating EigenTrust to the Redecentralized Web. This particularly includes the state of the art regarding EigenTrust and migrating Trust Models to the Web. The demonstration of feasibility with an implementation prototype of the concept is part of this thesis as well as a suitable evaluation with exemplary and empirical use cases.
