Master theses

Optimizing query evaluation over distributed data pods

Keywords: Linked Data, decentralized network, query evaluation, query optimization

Supervision: Pieter Colpaert

Students: max 1

To combat the centralization of data in large hubs on the Web such as Facebook and Google, technologies are being developed to enable Web Services to provide such functionality over decentralized data stores (pods). In order to provide both performance and data completeness when querying data over large amounts of decentralized data pods, optimizing the query evaluation to make use of present metadata is an important step in the query process.

The objective of this thesis is to contribute to optimizing the query evaluation process over decentralized data pods by incorporating available indexing metadata and data shape information in the query process. This research will help us create better querying solutions for decentralized data pods.

Your task is to:

  • Automate link extraction based on index structure metadata and data shapes.
  • Develop a link relevance scoring algorithm for the extracted hypermedia links.
  • Develop an implementation of your work in the Comunica framework.
  • Possibly assessing the scaling limitations of the approach, discussing approaches for data caching in the network to speedup the process