In this section, we show how our software infrastructure can be
used to implement two data-intensive applications—the K-mer Matching Interface (KMI)
4
and Breadth First Search (BFS), and compare
their performance to the respective MPI implementations. KMI is a
distributed, domain-specific library for hashing, designed for genome
assembly and mapping analysis applications. BFS is a fundamental
kernel in graph analytics, that is widely employed in many graph algorithms, including connected components, betweenness centrality,
and heuristic search algorithms [27,30].