[Profile picture of Ruben Verborgh]

Ruben Verborgh

Optimizing Approximate Membership Metadata in Triple Pattern Fragments for Clients and Servers

Ruben Taelman, Joachim Van Herwegen, Miel Vander Sande, and Ruben Verborgh

Depending on the HTTP interface used for publishing Linked Data, the effort of evaluating a SPARQL query can be redistributed differently between clients and servers. For instance, lower server-side CPU usage can be realized at the expense of higher bandwidth consumption. Previous work has shown that complementing lightweight interfaces such as Triple Pattern Fragments (TPF) with additional metadata can positively impact the performance of clients and servers. Specifically, Approximate Membership Filters (AMFs)—data structures that are small and probabilistic—in the context of TPF were shown to reduce the number of HTTP requests, at the expense of increasing query execution times. In order to mitigate this significant drawback, we have investigated unexplored aspects of AMFs as metadata on TPF interfaces. In this article, we introduce and evaluate alternative approaches for server-side publication and client-side consumption of AMFs within TPF to achieve faster query execution, while maintaining low server-side effort. Our alternative client-side algorithm and the proposed server configurations significantly reduce both the number of HTTP requests and query execution time, with only a small increase in server load, thereby mitigating the major bottleneck of AMFs within TPF. Compared to regular TPF, average query execution is more than 2 times faster and requires only 10% of the number of HTTP requests, at the cost of at most a 10% increase in server load. These findings translate into a set of concrete guidelines for data publishers on how to configure AMF metadata on their servers.

full text BibTeX other citation formats

Published in 2020 in Proceedings of the 13th International Workshop on Scalable Semantic Web Knowledge Base Systems.

Keywords:

Read this article online

Cite this article in your work

Cite this article easily using its BibTeX entry:

@inproceedings{taelman_swss_2020,
  author = {Taelman, Ruben and Van Herwegen, Joachim and Vander Sande, Miel and Verborgh, Ruben},
  title = {Optimizing Approximate Membership Metadata in {Triple Pattern Fragments} for Clients and Servers},
  booktitle = {Proceedings of the 13th International Workshop on Scalable Semantic Web Knowledge Base Systems},
  year = 2020,
  month = nov,
  series = {CEUR Workshop Proceedings},
  volume = 2757,
  issn = {1613-0073},
  pages = {1--16},
  url = {https://comunica.github.io/Article-SSWS2020-AMF/},
}

Alternatively, pick a reference of your choice below:

ACM
Ruben Taelman, Joachim Van Herwegen, Miel Vander Sande, and Ruben Verborgh. 2020. Optimizing Approximate Membership Metadata in Triple Pattern Fragments for Clients and Servers. In Proceedings of the 13th International Workshop on Scalable Semantic Web Knowledge Base Systems (CEUR Workshop Proceedings), 1–16.
APA
Taelman, R., Van Herwegen, J., Vander Sande, M., & Verborgh, R. (2020). Optimizing Approximate Membership Metadata in Triple Pattern Fragments for Clients and Servers. Proceedings of the 13th International Workshop on Scalable Semantic Web Knowledge Base Systems, 2757, 1–16.
IEEE
R. Taelman, J. Van Herwegen, M. Vander Sande, and R. Verborgh, “Optimizing Approximate Membership Metadata in Triple Pattern Fragments for Clients and Servers,” in Proceedings of the 13th International Workshop on Scalable Semantic Web Knowledge Base Systems, 2020, vol. 2757, pp. 1–16.
LNCS
Taelman, R., Van Herwegen, J., Vander Sande, M., Verborgh, R.: Optimizing Approximate Membership Metadata in Triple Pattern Fragments for Clients and Servers. In: Proceedings of the 13th International Workshop on Scalable Semantic Web Knowledge Base Systems. pp. 1–16 (2020).
MLA
Taelman, Ruben, et al. “Optimizing Approximate Membership Metadata in Triple Pattern Fragments for Clients and Servers.” Proceedings of the 13th International Workshop on Scalable Semantic Web Knowledge Base Systems, vol. 2757, 2020, pp. 1–16.

Discuss this article