ABSTRACTS OF ARTICLES OF THE JOURNAL "INFORMATION TECHNOLOGIES".
No. 7. Vol. 24. 2018

DOI: 10.17587/it.24.464-474

S. V. Gaevoy, Ph. D. (Engineering), Senior Lecturer, e-mail: gaevserge@mail.ru; W. M. A. Ahmed, Post-Graduate Student, e-mail: wesamalsofi@gmail.com; S. A. Fomenkov, D. Sc. (Engineering), Professor, Professor of "CAD" department, e-mail: saf@vstu.ru, Volgograd State Technical University, Volgograd, 400005, Russian Federation

Using Simulation to Determine Quality of Service of a Cluster System

In this paper computing clusters (CC) are considered. They are used to execute incoming jobs. There is such a CC in our university and we need to predict its service characteristics at executing several workloads. An important method to analyze parallel workloads is modeling execution of those systems by using parallel workload models (PWM). A simulation model of cluster has already been created at our department. We have already proposed many PWMs, but all these PWMs use a continuous variable approximation. This approximation can be done either by method of moments (MM), or maximum likelihood method (MLM). The latter gives the more accurate results but consumes much time. It was empirically proved in our and third-party papers. In this paper the goal is to demonstrate the already built CC model and the basics of its implementation. The implementation uses software design patterns and can adapted to many computing systems. We have chosen Java language because it has good performance and is convenient for the development. The program is cross-platform. Everyone can use, for example, Open JDK to compile it for free. The parallel workload of a cluster UniLu-GAIA is used because it is in free access and thus the computing can be repeated by everyone (although the nature of the tasks being executed there is confidential and unknown). The log being used belongs to the cluster with width 2004 at the moment of executing.
Keywords: parallel workload, parallel workload model, rigid job, job length, simulation, stochastic approximation, quality of service, computing cluster, simulation utility, workload approximation, continuous variable distribution, design pattern

P. 464–474

To the contents