The food supply chain network is a complex system involving various subsystems such as stock management, feed harvesting, cold storage, transportation, retail businesses, and regulatory certifications for food production. Throughout the food supply chain, major subsystems are owned by private organizations that share either little, or no information with other organizations. The restricted information flow, due to the fragmented and disjoint nature of the supply chain, results in reduced trust and traceability. With no knowledge shared, from the sequence of processes at different stages of the supply chain, an opportunity is lost to optimize the chain for better economic and environmental outcomes. The current technology in place for data communication, which relies on private and centralized ledgers, does not facilitate in the dissemination of critical information across the food supply chain network. This form of technology further limits the ability to collaborate on traceability, knowledge transfer and federated machine learning applications, because different subsets of the common data are owned by different private entities. In this thesis, we propose a decentralized and distributed supply chain connectivity and collaboration framework that is paired with blockchain technology, distributed resources, application and methods to enable reliable and non-pervasive food supply chain pertinent data consumption, data management, information extraction and knowledge transfer in a collaborative way. The proposed framework facilitates timely dissemination of critical information that is common to collaborating organizations, without any concerns for privacy, security and loss of data control. As a result of the helpful information dissemination from end-to-end, trust, transparency, traceability and collaboration are promoted.The technical contribution of this thesis lies in the generic, scalable, decentralized and distributed user controlled framework, that allows extracting and utilizing vital information from organizational data at different levels of the supply chain, along with its dissemination from end-to-end without any concerns of privacy, security, immutability and loss of data ownership. Seamless configuration and integration of distributed application and resources to support and enable, reliable federated machine learning data pipelines, makes the framework ideal for collaboration among distributed and disjoint organizations in the food supply chain network. Taking into account complex supply chains, the proposed extensible connectivity and collaboration framework allows integrating major types of information sources, (for example streaming data, hybrid databases, data feeds and static data sinks), while ensuring reliable and tamper-proof traceability, as data flows through collaboration channels. Information in the proposed framework is extracted and securely propagated using an integrated hierarchical blockchain infrastructure, coupled with distributed data storage, that is configured in a private network setting according to the organizational layout of participating supply chain actors. The organization controlled communication channels that are enabled in ad-hoc scenarios for collaboration, facilitate participants to communicate policies and decisions along with the implementation of numerous useful supply chain applications. Examples of some food supply chain related policies and applications that can be implemented by our proposed framework include, trading of carbon credits, tracking cattle in beef supply chain, jointly managing greenhouse gas emissions and optimizing end-to-end supply chain resource consumption. Strategies and techniques for protecting and securing the proposed distributed blockchain-based framework from the viewpoint of user accessibility, data integrity, confidentiality and privacy are also incorporated. The proposed framework takes into account, considerations for detailed software application level security measures to further enhance user trust. By incorporating a `Beef Supply Chain' example scenario, evaluation of an implemented application (named BeefMesh) has been done to prove its efficacy for collaboration, policy sharing, traceability, secure federated learning architectures, knowledge transfer and increased value for supply chain participants.
Read
