HyperProv: Decentralized Resilient Data Provenance at the Edge with Blockchains
Authors:
Petter Tunstad,
Amin M. Khan,
Phuong Hoai Ha
Abstract:
Data provenance and lineage are critical for ensuring integrity and reproducibility of information in research and application. This is particularly challenging for distributed scenarios, where data may be originating from decentralized sources without any central control by a single trusted entity. We present HyperProv, a general framework for data provenance based on the permissioned blockchain…
▽ More
Data provenance and lineage are critical for ensuring integrity and reproducibility of information in research and application. This is particularly challenging for distributed scenarios, where data may be originating from decentralized sources without any central control by a single trusted entity. We present HyperProv, a general framework for data provenance based on the permissioned blockchain Hyperledger Fabric (HLF), and to the best of our knowledge, the first system that is ported to ARM based devices such as Raspberry Pi (RPi). HyperProv tracks the metadata, operation history and data lineage through a set of built-in queries using smart contracts, enabling lightweight retrieval of provenance data. HyperProv provides convenient integration through a NodeJS client library, and also includes off-chain storage through the SSH file system. We evaluate HyperProv's performance, throughput, resource consumption, and energy efficiency on x86-64 machines, as well as on RPi devices for IoT use cases at the edge.
△ Less
Submitted 13 October, 2019;
originally announced October 2019.
