Year: 2024
Type: conference-paper
Abstract: In the digital era, supply chains have evolved into complex, Internet of Things (IoT)-driven networks requiring efficient data sharing and processing. Traditional data-sharing methods face challenges in the current predictive supply chain's need for processed data as insights are required in near-real-time. However, existing messaging models often suffer from the lack of support for automated processes, or the centralisation issues, such as single points of failure and potential bottlenecks. This paper introduces a Decentralised Publish-Process-Subscribe (DPPS) middleware to overcome these limitations, offering byzantine fault-tolerant data processing and verifiable insights without relying on a central entity. DPPS integrates data processing with the pub-sub paradigm, enabling automatic raw data processing in a decentralised manner. The paper presents a decentralised data sharing and processing architecture, a byzantine fault-tolerant verification method, and evaluates the proposed system through a simulated supply chain scenario. Results indicate that DPPS achieves decentralised, verifiable insights with no transaction fees and minimal latency trade-offs, addressing the shortcomings of existing messaging middlewares in predictive supply chains.
Authors: Amir Jabbari, Gowri Ramachandran, Sidra Malik, Raja Jurdak
Institution: Queensland University of Technology
Language: English
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W4402594169