A Universal Graph Transaction-Log
Luis Garcés-Erice, Enrico Toniato, Daniel Bauer, Peter Urbanetz, Seán Rooney · 2024
We demonstrate the use of a transaction-log that maintains a graph in a serialized form and show how this graph can be materialized in multiple different graph processing systems for specific use-cases. Our demonstration uses this log to build a data observability framework for enterprise-scale data processing systems. In the demo we show two different graph materializations: one generating reports for compliance officers, the other allowing data scientists to perform analytics. We demonstrate the greater flexibility of our approach over creating distinct views within the same graph database.