Secure Distributed Computing Managed by Blockchain

Abstract

Moving large amounts of data between networks for data analysis and computations presents several issues related to privacy and security. In collaboration with the TOTEM project [1], we propose a solution to these problems, by moving computations to the residence of the data. We introduce a novel approach for managing access to remote datasets and resources by blockchain technology through Hyperledger Fabric. Organizations with similar interests may join a consortium, which will form a private channel on the blockchain network, i.e., a separate ledger. Participating organizations will enroll their users, who thereafter must obtain a one-time-code using a smart contract in order to gain access to remote resources. We utilize Ansible for remotely deploying Hadoop clusters for computation, which will comprise several Docker containers. A user may run computations at several remote locations separately, and subsequently retrieve a combined result without having to share data between organizations. To ensure privacy between participating organizations we utilize chaincode and private data collections in Hyperledger Fabric. Finally, we demonstrate three ways of deploying the solution: locally, as a single cluster in the cloud using Azure, and across multiple clusters in the cloud using Azure. Our solution ensures data privacy by allowing data providing organizations to connect their own computational resources for data consumers to use. By running computations inside Docker containers on these resources, we ensure that these processes are isolated from the host system.