Towards the use of semantic learning object repositories: Evaluating queries performance in two different RDF implementations

Learning Object Repositories (LOR) are an essential component of the e-Learning ecosystem and have being normally serving the purpose of cataloging, storing, retrieving and delivering Learning Objects to be used inside e-Learning applications. Next generation of LORs needs to overcome some major shortcomings current LORs present and involve other entities that are part of the e-learning process (teachers, students, lessons, courses, activities, learning paths, etc.) in a way that they are all fully integrated and linked-up. Semantic web technologies such as RDF are the natural choice to implement these requirements into Semantic Learning Repositories. One important factor limiting the implementation of this kind of systems is the uncertainty about their performance. The present paper describes an initial study that compares the performance of two distinct RDF native database implementations (4store and Jena Apache) in the specific context of a Semantic Learning Repository. The performance tests were run to evaluate two different aspects of the databases implementation: the time to upload the RDF data to the databases, and the response time for running the queries. The results showed that 4store performed better than Apache Jena for all the scenarios we evaluated.