Cloud-Based RDF Data Management That'S Both Powerful And Extensible

Abstract

Even if there have been some recent improvements in the administration of distributed RDF data, it is still rather difficult to do analysis on large amounts of RDF data using the cloud. Although having a very easy data paradigm, RDF is capable of storing complex graphs that mix information at the instance-level and the schema-level. The distributed operations that are produced as a consequence of sharding this sort of data using standard approaches, such as partitioning the graph using usual min-cut algorithms, are exceedingly inefficient and call for a number of joins to be performed. In this paper, we explore DC, a cloud-optimized distributed RDF data management system that is both effective and scalable. It was created primarily for use in cloud environments. In contrast to more conventional approaches, DC first does a physiological analysis on both the instance data and the schema data before it divides the data. In this paper, we provide an overview of the architecture of DC, covering its fundamental data structures as well as the innovative approaches that we use for the division and distribution of data. In addition to this, we provide a comprehensive analysis of DC, which demonstrates that, for the vast majority of workloads, our system is often twice as fast as the most modern alternatives.