Publication
DSN-Industry Track 2019
Conference paper

System Restore in a Multi-cloud Data Pipeline Platform

View publication

Abstract

Data pipeline platforms hosting big data analytics can span multiple clouds. Backup and restore service is typically applied to deal with data corruptions in such platforms. This paper proposes a novel approach to providing consistency to the restored state of a multi-cloud data pipeline platform from its backups, and also presents the performance of the approach demonstrated in a dry run test.