Efficient Semi-External Depth-First Search
share
Computing Depth-First Search (DFS) results, i.e. depth-first order or DFS-Tree, on the semi-external environment becomes a hot topic, because the scales of the graphs grow rapidly which can hardly be hold in the main memory, in the big data era. Existing semi-external DFS algorithms assume the main memory could, at least, hold a spanning tree T of a graph G, and gradually restructure T into a DFS-Tree, which is non-trivial. In this paper, we present a comprehensive study of semi-external DFS problem, including the first theoretical analysis of the main challenge of this problem, as far as we know. Besides, we introduce a new semi-external DFS algorithm with an efficient edge pruning principle, named EP-DFS. Unlike the traditional algorithms, we not only focus on addressing such complex problem efficiently with less I/Os, but also focus on that with simpler CPU calculation (Implementation-friendly) and less random I/O access (key-to-efficiency). The former is based on our efficient pruning principle; the latter is addressed by a lightweight index N+-index, which is a compressed storage for a subset of the edges for G. The extensive experimental evaluation on both synthetic and real graphs confirms that our EP-DFS algorithm outperforms the existing techniques.
READ FULL TEXT
