fediscount-website

Fediscount is a proof of concept for Web applications relying on configurable federations of SPARQL endpoints. It is an e-commerce Website where end users can browse a set of products and offers made available by a set of independent vendors.

       

Fediscount comprises 3 components: (i) a Web view accessible through modern Web browsers, (ii) a federation engine able to process queries for large scale federations in reasonable time, and (iii) a federation of SPARQL endpoints.

A Web View

This repository hosts the view written using Vue.js 3. In this spirit, every view is a component; and when needed, it communicates its specific queries to the Log that calls the federations engine and dispatches the results upon receipt.

For a quick deployment, please make sure that node.js is installed, and follow the instructions:

git clone https://github.com/GDD-Nantes/fediscount-website.git
cd fediscount-website
npm install -D

# Run a vite server listening on 5173 by default
npm run dev

A Federation Engine

Not all federation engines are fit to answer large scale federation of endpoints [1]. To ensure reasonable response times, we use FedUP [2], a federation engine that, compared to the state-of-the-art, builds better logical plans, among others favoring the identification of exclusive groups.

To deploy FedUP for Fediscount, one must first create 3 tiny quotient summaries of the targeted FedShop federations (20 endpoints, 100 endpoints, and 200 endpoints) using Apache Jena.

Note

You can download the 3 summaries on the Zenodo platform.

Once created, please run the HTTP server by executing the command as follows:

git clone https://github.com/GDD-Nantes/fedup
cd fedup

# Run the federation engine server on port 3330 by default
# Replace each summary path by yours
mvn exec:java -Dexec.args="--summaries=./fedshop100-h0,./fedshop20-h0,./fedshop200-h0 --engine=FedX --export"

A SPARQL Federation

Large scale federations are infamously tricky to deploy. To alleviate this issue, we emulate a federation of 200 endpoints using a single Virtuoso endpoint hosting multiple graphs. We used Version 07.20.3233-pthreads for Mac OS 11 (Apple Silicon) as of Jun 22 2021, and ingested FedShop's federation of 200 endpoints with a .ini config file exposing the API on port 5555.

Note

You can download the .db database file on the Zenodo platform.

References

[1] Minh-Hoang Dang, Julien Aimonier-Davat, Pascal Molli, Olaf Hartig, Hala Skaf-Molli, and Yotlan Le Crom. FedShop: A Benchmark for Testing the Scalability of SPARQL Federation Engines. In The 22th International Semantic Web Conference (ISWC), Athens, Greece, November, 2023.

[2] Julien Aimonier-Davat, Minh-Hoang Dang, Pascal Molli, Brice Nédelec, and Hala Skaf-Molli. FedUP: Querying Large-Scale Federations of SPARQL Endpoints. In The Web Conference 2024 (WWW’2024). Singapore, 2024.

[3] Julien Aimonier-Davat, Minh-Hoang Dang, Pascal Molli, Brice Nédelec, Hala Skaf-Molli. Fediscount: Shopping Online at a Federated Store Using FedUP as SPARQL Federation Engine. The ACM Web Conference (WWW ’24), May 2024, Singapore.