3D Inverted Index with Cache Sharing for Web Search Engines

ESTEBAN FEUERSTEIN; GIL COSTA GRACIELA VERONICA; MAURICIO MARIN; GABRIEL TOLOSA; RICARDO BAEZA-YATES

European Conference on Parallel and Distributed Computing

Springer-Verlag Berlin Heidelberg

Año: 2012; p. 272 - 282

Web search engines achieve efficient performance by partitioningand replicating the indexing data structure used to support queryprocessing. Current practice simply partitions and replicates the textcollection on the set of cluster processors and then constructs in eachprocessor an index data structure. This paper proposes a different approachby constructing an index data structure that properly considersthe fact that data is partitioned and replicated. This leads to a so-called3D indexing strategy that outperforms current approaches. Performanceis further boosted by introducing an application caching scheme devisedto hold most frequently issued queries.