Adaptive delta compression over versioned knowledge graphs
Keywords: adaptive, archiving, compression, delta, knowledge graph, rdf, versioning
Promotors: Ruben Verborgh, Ruben Taelman
Students: max 1
Problem
TerminusDB is an open source knowledge graph and document store.
It allows multiple versions of a knowledge graph to be stored efficiently.
It uses an efficient internal format form storing a snapshot for the initial version, and deltas containing relative additions and deletions for every next version.
In some cases, the stored deltas contain a lot of redundancy, which makes storage efficiency sub-optimal.
In these cases, it may be relevant to compress multiple deltas together.
Goal
In this thesis, you will investigate how multiple deltas can be compressed together efficiently, and in what cases this delta compression should occur.
You will collaborate with the team of developers at TerminusX and researchers at Ghent University.