A Tensor-Based Framework for Software-Defined Cloud Data Center

Liwei Kuang, Laurence T. Yang, Seungmin (Charlie) Rho, Zheng Yan, Kai Qiu

    Research output: Contribution to journalArticleScientificpeer-review

    7 Citations (Scopus)


    Multimedia has been exponentially increasing as the biggest big data, which consist of video clips, images, and audio files. Processing and analyzing them on a cloud data center have become a preferred solution that can utilize the large pool of cloud resources to address the problems caused by the tremendous amount of unstructured multimedia data. However, there exist many challenges in processing multimedia big data on a cloud data center, such as multimedia data representation approach, an efficient networking model, and an estimation method for traffic patterns. The primary purpose of this article is to develop a novel tensor-based software-defined networking model on a cloud data center for multimedia big-data computation and communication. First, an overview of the proposed framework is provided, in which the functions of the representative modules are briefly illustrated. Then, three models,-forwarding tensor, control tensor, and transition tensor-are proposed for management of networking devices and prediction of network traffic patterns. Finally, two algorithms about single-mode and multimode tensor eigen-decomposition are developed, and the incremental method is employed for efficiently updating the generated eigen-vector and eigen-tensor. Experimental results reveal that the proposed framework is feasible and efficient to handle multimedia big data on a cloud data center.
    Original languageEnglish
    Article number2983640
    Number of pages23
    Issue number5s
    Publication statusPublished - 1 Dec 2016
    MoE publication typeA1 Journal article-refereed


    • big data
    • tensor
    • software defined networks
    • data center


    Dive into the research topics of 'A Tensor-Based Framework for Software-Defined Cloud Data Center'. Together they form a unique fingerprint.

    Cite this