Mobile and Ubiquitous Systems: Computing, Networking, and Services. 10th International Conference, MOBIQUITOUS 2013, Tokyo, Japan, December 2-4, 2013, Revised Selected Papers

Research Article

Highly Distributable Associative Memory Based Computational Framework for Parallel Data Processing in Cloud

Download
539 downloads
  • @INPROCEEDINGS{10.1007/978-3-319-11569-6_6,
        author={Amir Basirat and Asad Khan and Balasubramaniam Srinivasan},
        title={Highly Distributable Associative Memory Based Computational Framework for Parallel Data Processing in Cloud},
        proceedings={Mobile and Ubiquitous Systems: Computing, Networking, and Services. 10th International Conference, MOBIQUITOUS 2013, Tokyo, Japan, December 2-4, 2013,  Revised Selected Papers},
        proceedings_a={MOBIQUITOUS},
        year={2014},
        month={12},
        keywords={Associative memory Neural networks Big data MapReduce Graph Neuron},
        doi={10.1007/978-3-319-11569-6_6}
    }
    
  • Amir Basirat
    Asad Khan
    Balasubramaniam Srinivasan
    Year: 2014
    Highly Distributable Associative Memory Based Computational Framework for Parallel Data Processing in Cloud
    MOBIQUITOUS
    Springer
    DOI: 10.1007/978-3-319-11569-6_6
Amir Basirat1,*, Asad Khan1,*, Balasubramaniam Srinivasan1,*
  • 1: Monash University Melbourne
*Contact email: Amir.Basirat@monash.edu, Asad.Khan@monash.edu, Bala.Srinivasan@monash.edu

Abstract

One of the main challenges for large-scale computer clouds dealing with massive real-time data is in coping with the rate at which unprocessed data is being accumulated. In this regard, associative memory concepts open a new pathway for accessing data in a highly distributed environment that will facilitate a parallel-distributed computational model to automatically adapt to the dynamic data environment for optimized performance. With this in mind, this paper targets a new type of data processing approach that will efficiently partition and distribute data for clouds, providing a parallel data access scheme that enables data storage and retrieval by association where data records are treated as patterns; hence, finding overarching relationships among distributed data sets becomes easier for a variety of pattern recognition and data-mining applications. The ability to partition data optimally and automatically will allow elastic scaling of system resources and remove one of the main obstacles in provisioning data centric software-as-a-service in clouds.