Eblocbroker: An autonomous blockchain-based computational broker for e-science

Blockchain infrastructures have emerged as disruptive technology, enabling cryptocurrencies and smart contracts. They can have a wide range of application areas in e-Science due to their open, public nature and global accessibility in a trustless manner. We propose and implement eBlocBroker, an autonomous blockchain-based middleware for selling computing power and providing data resources for eScience. eBlocBroker connects requesters with providers via an Ethereum-based private blockchain (Bloxberg) and uses cloud storage like B2DROP, IPFS, or Google Drive. In this thesis, we demonstrate how a smart contract can deploy a marketplace for data and computation-intensive research. We propose a cost model in the smart contract to calculate and record computation, transfer, storage, and cache usage costs. A Pythonbased system communicates with eBlocBroker and orchestrates job execution on the provider's end. We present eBlocBroker's features, infrastructure, implementation, algorithms, and experimental results. We also contribute an Ethereum blockchain-based scientific workflow execution manager that distributes workflows to cluster computing providers. We extended eBlocBroker for distributed workflow execution via blockchain, demonstrating its effectiveness through various tests. Finally, we introduce a decentralized autonomous software organization model and the AutonomousSoftwareOrg smart contract, providing project funding, voting mechanisms, and recordkeeping for software usage. Furthermore, software executions and their data files can be transactionally recorded, enabling the construction of execution graphs for analysis. A model of input/output data and software executions is presented using an AND/OR graph.