A taxonomy of scheduling in general-purpose distributed computing systems

Abstract

One measure of usefulness of a general-purpose distributed computing system is the system's ability to provide a level of performance commensurate to the degree of multiplicity of resources present in the system. Many different approaches and metrics of performance have been proposed in an attempt to achieve this goal in existing systems. In addition, analogous problem formulations exist in other fields such as control theory, operations research, and production management. However, due to the wide variety of approaches to this problem, it is difficult to meaningfully compare different systems since there is no uniform means for qualitatively or quantitatively evaluating them. It is difficult to successfully build upon existing work or identify areas worthy of additional effort without some understanding of the relationships between past efforts. In this paper, a taxonomy of approaches to the resource management problem is presented in an attempt to provide a common terminology and classification mechanism necessary in addressing this problem. The taxonomy, while presented and discussed in terms of distributed scheduling, is also applicable to most types of resource management. As an illustration of the usefulness of the taxonomy an annotated bibliography is given which classifies a large number of distributed scheduling approaches according to the taxonomy.

On Load Balancing in Distributed Telecommunication Systems

“…Using the classification of Casavant and Kuhl, 1988 (see Figure 2), we learn from the so far sparse literature on LB in telecom systems ( Kihl et aI., 1999;Rumeswicz, 1999;Arvidsson et aI., 1999;Osland and Emstad, 1999), that dynamic, distributed load balancing policies seem to be best suited. A dynamic (or adaptive) policy takes system state into account when scheduling new service requests.…”

Teaching about Madrid: A Collaborative Agents-Based Distributed Learning Course

“…A taxonomy of load balancing methods has been defined in [3], taking into account different aspects. Three important criteria for this classification are: Time in which workload distribution is performed static [6] or dynamic [11]; Control which can be centralized [10] or distributed [6] and System state view global [6] or local [4].…”

A Better Performance to INPE Satellite Control Software

“…MALBA architecture belongs to the dynamic distributed classification described in Ref. 8, which consists of making balancing decisions at run time and the work involved in making these decisions is physically distributed among the nodes. This descentralized approach makes MALBA a safe load balancing service because there is not the existence of a unique failure point in the system whereas non-distributed load balancing approaches 9, 7, 10 might put load balancing in risk if the central node fails.…”

“…Task migration could be either preemptive or nonpreemptive 8 . Preemptive task migration allows tasks in execution to be stopped, transferred, and then resumed at another node.…”