The thesis addresses trust management in the pharmaceutical supply
chain, where multiple independent actors exchange data, certificates and
other information sources. It proposes an architecture in which an ontology,
verifiable credentials and a smart contract on a blockchain jointly form
a trust graph. The key element is a decentralized oracle network, where
each oracle applies its own trust rules. Oracles consume heterogeneous data
sources to compute a binary trust decision and a numerical trust score, and
record these results in a smart contract on an Ethereum test network. The
prototype is implemented using Docker containers and multiple oracles with
different policies, and is evaluated on repeatable scenarios that show how
trust scores change as input data deteriorate or improve. Finally, the thesis
presents a STRIDE-based security analysis of the smart contract and oracle
network and outlines the enhancements required for deployment in a
production environment and for broader use in other domains.
|
