FR-index: A multi-dimensional indexing framework for switch-centric data centers

Abstract

Data center occupies a decisive position in business of data management and data analysis. To improve the efficiency of data retrieval in a data center, we propose a distributed multi-dimensional indexing framework for switch-centric data centers with tree-like topologies. Taking Fat-Tree as a representative, which is a typical switch-centric data center topology, we design FR-Index, a two-layer indexing schema fully taking advantage of the Fat-Tree topology and R-tree indexing technology. In the lower layer, each server indexes the local data with R-tree, while in the upper layer the distributed global index depicting an overview of the whole data set. To improve the efficiency of query processing, we also provide special techniques to reduce the dimensionality of the index. Experiments on Amazon’s EC2 show that our proposed indexing schema is scalable, efficient and lightweight, which can significantly promote the efficiency of query processing.