Optimal load balancing and scheduling in a distributed computer system

Keith Ross, David D. Yao

    Research output: Contribution to journalArticle

    Abstract

    A broadcast communication network that interconnects multiple hosts is considered. A job generated by a host is either generic, in which case it can be processed at any of the hosts in the network, or dedicated, in which case it must be processed at the host from which it originates. It is assumed that each host can generate several types of dedicated jobs, where each type imposes a constraint on its average response time. The optimal design objective is twofold: First, to redistribute the generic jobs among the host computers in order to minimize the average response time, which shall be referred to as the load-balancing problem; second, for each host computer, to design a control that schedules the dedicated and the generic jobs so that the response-time constraints are met for each of the dedicated traffic types. Despite its complexity, the model described above has some attractive features. Specifically, for a given allocation of the generic traffic, the scheduling problem at each host can be solved as a polymatroid optimization problem. The underlying polymatroid structure leads to an efficient algorithm to determine, as a function of the offered load of generic traffic, the average delay of the generic jobs at any given host. Further analysis reveals that the delay functions at each of the hosts are convex and increasing. Therefore, an allocation algorithm can be employed to solve the system-level load-balancing problem. An example is provided indicating that a substantial improvement in performance can be obtained by incorporating scheduling into the load-balancing procedure.

    Original languageEnglish (US)
    Pages (from-to)676-690
    Number of pages15
    JournalJournal of the ACM
    Volume38
    Issue number3
    StatePublished - 1991

    Fingerprint

    Distributed computer systems
    Resource allocation
    Scheduling
    Telecommunication networks

    Keywords

    • Algorithms
    • Load balancing
    • Management
    • Performance
    • Queuing theory
    • Scheduling

    ASJC Scopus subject areas

    • Artificial Intelligence
    • Hardware and Architecture
    • Information Systems
    • Software
    • Control and Systems Engineering

    Cite this

    Optimal load balancing and scheduling in a distributed computer system. / Ross, Keith; Yao, David D.

    In: Journal of the ACM, Vol. 38, No. 3, 1991, p. 676-690.

    Research output: Contribution to journalArticle

    Ross, Keith ; Yao, David D. / Optimal load balancing and scheduling in a distributed computer system. In: Journal of the ACM. 1991 ; Vol. 38, No. 3. pp. 676-690.
    @article{6d45f25baf594d9c831819037d4c6cb7,
    title = "Optimal load balancing and scheduling in a distributed computer system",
    abstract = "A broadcast communication network that interconnects multiple hosts is considered. A job generated by a host is either generic, in which case it can be processed at any of the hosts in the network, or dedicated, in which case it must be processed at the host from which it originates. It is assumed that each host can generate several types of dedicated jobs, where each type imposes a constraint on its average response time. The optimal design objective is twofold: First, to redistribute the generic jobs among the host computers in order to minimize the average response time, which shall be referred to as the load-balancing problem; second, for each host computer, to design a control that schedules the dedicated and the generic jobs so that the response-time constraints are met for each of the dedicated traffic types. Despite its complexity, the model described above has some attractive features. Specifically, for a given allocation of the generic traffic, the scheduling problem at each host can be solved as a polymatroid optimization problem. The underlying polymatroid structure leads to an efficient algorithm to determine, as a function of the offered load of generic traffic, the average delay of the generic jobs at any given host. Further analysis reveals that the delay functions at each of the hosts are convex and increasing. Therefore, an allocation algorithm can be employed to solve the system-level load-balancing problem. An example is provided indicating that a substantial improvement in performance can be obtained by incorporating scheduling into the load-balancing procedure.",
    keywords = "Algorithms, Load balancing, Management, Performance, Queuing theory, Scheduling",
    author = "Keith Ross and Yao, {David D.}",
    year = "1991",
    language = "English (US)",
    volume = "38",
    pages = "676--690",
    journal = "Journal of the ACM",
    issn = "0004-5411",
    publisher = "Association for Computing Machinery (ACM)",
    number = "3",

    }

    TY - JOUR

    T1 - Optimal load balancing and scheduling in a distributed computer system

    AU - Ross, Keith

    AU - Yao, David D.

    PY - 1991

    Y1 - 1991

    N2 - A broadcast communication network that interconnects multiple hosts is considered. A job generated by a host is either generic, in which case it can be processed at any of the hosts in the network, or dedicated, in which case it must be processed at the host from which it originates. It is assumed that each host can generate several types of dedicated jobs, where each type imposes a constraint on its average response time. The optimal design objective is twofold: First, to redistribute the generic jobs among the host computers in order to minimize the average response time, which shall be referred to as the load-balancing problem; second, for each host computer, to design a control that schedules the dedicated and the generic jobs so that the response-time constraints are met for each of the dedicated traffic types. Despite its complexity, the model described above has some attractive features. Specifically, for a given allocation of the generic traffic, the scheduling problem at each host can be solved as a polymatroid optimization problem. The underlying polymatroid structure leads to an efficient algorithm to determine, as a function of the offered load of generic traffic, the average delay of the generic jobs at any given host. Further analysis reveals that the delay functions at each of the hosts are convex and increasing. Therefore, an allocation algorithm can be employed to solve the system-level load-balancing problem. An example is provided indicating that a substantial improvement in performance can be obtained by incorporating scheduling into the load-balancing procedure.

    AB - A broadcast communication network that interconnects multiple hosts is considered. A job generated by a host is either generic, in which case it can be processed at any of the hosts in the network, or dedicated, in which case it must be processed at the host from which it originates. It is assumed that each host can generate several types of dedicated jobs, where each type imposes a constraint on its average response time. The optimal design objective is twofold: First, to redistribute the generic jobs among the host computers in order to minimize the average response time, which shall be referred to as the load-balancing problem; second, for each host computer, to design a control that schedules the dedicated and the generic jobs so that the response-time constraints are met for each of the dedicated traffic types. Despite its complexity, the model described above has some attractive features. Specifically, for a given allocation of the generic traffic, the scheduling problem at each host can be solved as a polymatroid optimization problem. The underlying polymatroid structure leads to an efficient algorithm to determine, as a function of the offered load of generic traffic, the average delay of the generic jobs at any given host. Further analysis reveals that the delay functions at each of the hosts are convex and increasing. Therefore, an allocation algorithm can be employed to solve the system-level load-balancing problem. An example is provided indicating that a substantial improvement in performance can be obtained by incorporating scheduling into the load-balancing procedure.

    KW - Algorithms

    KW - Load balancing

    KW - Management

    KW - Performance

    KW - Queuing theory

    KW - Scheduling

    UR - http://www.scopus.com/inward/record.url?scp=0009417707&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=0009417707&partnerID=8YFLogxK

    M3 - Article

    VL - 38

    SP - 676

    EP - 690

    JO - Journal of the ACM

    JF - Journal of the ACM

    SN - 0004-5411

    IS - 3

    ER -