How to store and query very large graphs is a problem that has been attracting much interest in the database research community. Currently, the mainstream approach to improve the scalability is to use distributed computing architectures, which offer many cores and storage space. However, the benefit of getting much more computing power comes at the price of higher communication cost and increased system complexity. Moreover, current solutions tend to focus on specific tasks, leaving the user no choice but to load the graph into different engines whenever the graph is needed for different tasks. These considerations lead us to one question: How far can non-distributed (centralised) solutions for graph processing go in terms of scalability and ability to serve multiple workloads? In this talk, I will try to answer this question by presenting Trident -- a new graph database designed for centralised architectures. Trident can store very large graphs by exploiting compression and by optimizing the storage depending on the graph topology. Moreover, Trident can support a wide range of workloads by providing a set of low-level and general-purpose primitives. In our experiments, we observed that Trident can load graphs with up to 10^11 (100 billions) edges using inexpensive commodity hardware, delivering state-of-the-art performance on tasks like SPARQL query answering or graph analytics.
Bio: Jacopo Urbani is a tenured assistant professor in Computer Science at the Vrije Universiteit Amsterdam (VUA). He is also a guest researcher at the Centrum Wiskunde & Informatica (CWI). He wrote a PhD thesis on distributed reasoning algorithms for very large Knowledge Graphs. The thesis was nominated by a committee from the Royal Netherlands Academy of Arts and Sciences as one of the best PhD theses in Computer Science in the country. After spending part of his postdoc in USA (Stanford) and Germany (MPII), he joined the faculty of the VUA and has been tenured in 2018. He leads a research unit on knowledge extraction and inference from large Web corpora which has developed tools that are considered among the state-of-the-art in the respective fields. For more information, please see https://www.jacopourbani.it
