Towards efficient operation of internet data center networks: Joint data placement and flow control for cost optimization

Abdallah Khreishah, Jacob Chakareski, Ammar Gharaibeh, Issa Khalil, Ali Diabat

    Research output: Contribution to journalArticle

    Abstract

    The problem of cost-efficient operation of data center networks used to deliver file sharing services is studied. The aggregate costs are split into server-load-related and link-load-related shares. Thus, the problem of interest is formulated as one of joint data placement and flow control, and mixed integer-linear programming is used to compute the optimal solution. The high complexity of the latter motivated us to design two additional sets of strategies, based on data coding and heuristics, respectively. With coding, a distributed algorithm for the problem is developed. In the simulation experiments, carried out based on actual data center information, network topology and link cost, as well as electricity prices, the advantages of data coding, in particular in the context of multicast, and the impact of different factors such as the network topology and service popularity, on the total cost incurred by all considered strategies, are examined. Network coding with multicast is shown to provide cost savings in the order of 30–80%, depending on the specific context under consideration, relative to the other optimization strategies and heuristic methods examined in this work.

    Original languageEnglish (US)
    Pages (from-to)83-98
    Number of pages16
    JournalSimulation Modelling Practice and Theory
    Volume64
    DOIs
    StatePublished - Jan 1 2016

    Fingerprint

    Data Placement
    Cost Optimization
    Data Center
    Flow Control
    Data Flow
    Flow control
    Internet
    Costs
    Coding
    Multicast
    Network Topology
    Topology
    Mixed Integer Linear Programming
    Heuristic methods
    Network coding
    Network Coding
    Information services
    Heuristic Method
    Distributed Algorithms
    Electricity

    Keywords

    • Cost efficiency
    • Data placement
    • Internet datacenter networks

    ASJC Scopus subject areas

    • Software
    • Modeling and Simulation
    • Hardware and Architecture

    Cite this

    Towards efficient operation of internet data center networks : Joint data placement and flow control for cost optimization. / Khreishah, Abdallah; Chakareski, Jacob; Gharaibeh, Ammar; Khalil, Issa; Diabat, Ali.

    In: Simulation Modelling Practice and Theory, Vol. 64, 01.01.2016, p. 83-98.

    Research output: Contribution to journalArticle

    @article{be7dfbe80af746ea9668942b48413d51,
    title = "Towards efficient operation of internet data center networks: Joint data placement and flow control for cost optimization",
    abstract = "The problem of cost-efficient operation of data center networks used to deliver file sharing services is studied. The aggregate costs are split into server-load-related and link-load-related shares. Thus, the problem of interest is formulated as one of joint data placement and flow control, and mixed integer-linear programming is used to compute the optimal solution. The high complexity of the latter motivated us to design two additional sets of strategies, based on data coding and heuristics, respectively. With coding, a distributed algorithm for the problem is developed. In the simulation experiments, carried out based on actual data center information, network topology and link cost, as well as electricity prices, the advantages of data coding, in particular in the context of multicast, and the impact of different factors such as the network topology and service popularity, on the total cost incurred by all considered strategies, are examined. Network coding with multicast is shown to provide cost savings in the order of 30–80{\%}, depending on the specific context under consideration, relative to the other optimization strategies and heuristic methods examined in this work.",
    keywords = "Cost efficiency, Data placement, Internet datacenter networks",
    author = "Abdallah Khreishah and Jacob Chakareski and Ammar Gharaibeh and Issa Khalil and Ali Diabat",
    year = "2016",
    month = "1",
    day = "1",
    doi = "10.1016/j.simpat.2015.09.010",
    language = "English (US)",
    volume = "64",
    pages = "83--98",
    journal = "Simulation Modelling Practice and Theory",
    issn = "1569-190X",
    publisher = "Elsevier",

    }

    TY - JOUR

    T1 - Towards efficient operation of internet data center networks

    T2 - Joint data placement and flow control for cost optimization

    AU - Khreishah, Abdallah

    AU - Chakareski, Jacob

    AU - Gharaibeh, Ammar

    AU - Khalil, Issa

    AU - Diabat, Ali

    PY - 2016/1/1

    Y1 - 2016/1/1

    N2 - The problem of cost-efficient operation of data center networks used to deliver file sharing services is studied. The aggregate costs are split into server-load-related and link-load-related shares. Thus, the problem of interest is formulated as one of joint data placement and flow control, and mixed integer-linear programming is used to compute the optimal solution. The high complexity of the latter motivated us to design two additional sets of strategies, based on data coding and heuristics, respectively. With coding, a distributed algorithm for the problem is developed. In the simulation experiments, carried out based on actual data center information, network topology and link cost, as well as electricity prices, the advantages of data coding, in particular in the context of multicast, and the impact of different factors such as the network topology and service popularity, on the total cost incurred by all considered strategies, are examined. Network coding with multicast is shown to provide cost savings in the order of 30–80%, depending on the specific context under consideration, relative to the other optimization strategies and heuristic methods examined in this work.

    AB - The problem of cost-efficient operation of data center networks used to deliver file sharing services is studied. The aggregate costs are split into server-load-related and link-load-related shares. Thus, the problem of interest is formulated as one of joint data placement and flow control, and mixed integer-linear programming is used to compute the optimal solution. The high complexity of the latter motivated us to design two additional sets of strategies, based on data coding and heuristics, respectively. With coding, a distributed algorithm for the problem is developed. In the simulation experiments, carried out based on actual data center information, network topology and link cost, as well as electricity prices, the advantages of data coding, in particular in the context of multicast, and the impact of different factors such as the network topology and service popularity, on the total cost incurred by all considered strategies, are examined. Network coding with multicast is shown to provide cost savings in the order of 30–80%, depending on the specific context under consideration, relative to the other optimization strategies and heuristic methods examined in this work.

    KW - Cost efficiency

    KW - Data placement

    KW - Internet datacenter networks

    UR - http://www.scopus.com/inward/record.url?scp=84978734720&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=84978734720&partnerID=8YFLogxK

    U2 - 10.1016/j.simpat.2015.09.010

    DO - 10.1016/j.simpat.2015.09.010

    M3 - Article

    AN - SCOPUS:84978734720

    VL - 64

    SP - 83

    EP - 98

    JO - Simulation Modelling Practice and Theory

    JF - Simulation Modelling Practice and Theory

    SN - 1569-190X

    ER -