Stratified Multi-ring Distributed Search Model for Big Data
Weiqing Cheng, Geng Yang, Shanshan Zhang, Shaobai Zhang · 2014
Big data technologies still have a long way to go, the challenges including capture, content management, storage, search, sharing, transfer, analysis, and so on. Distributed Peer-to-Peer storage and search technologies due to their characteristics in distribution and scalability can be used for big data. Chord is a typical Distributed Hash Table-based lookup protocol used in P2P networks. To make it more suitable for big data environments and to promote its performance, this paper proposes a stratified multi-ring distributed search (SMRDS) model that consists of an upper core network, and lower multiple ring networks that are Chord-based and built concerning real network environment. In the model, nodes at different levels are managed differently, and search operations contain intra-regional and cross-regional queries. Experimental results show that the SMRDS has better search performance in term of the average search delay than Chord. Finally, this paper provides a big data application using the proposed model.