A book for experts and practitioners, emphasizing the intuition and reasoning behind definitions and derivations related to evaluating computer systems performance.
Based on the author's experience in industry, this book focuses on simple techniques for solving everyday problems in systems design and analysis. All techniques are covered in a non-mathematical way, so that no statistics expertise is necessary.
The only singular, all-encompassing textbook on state-of-the-art technical performance evaluation Fundamentals of Performance Evaluation of Computer and Telecommunication Systems uniquely presents all techniques of performance evaluation of computers systems, communication networks, and telecommunications in a balanced manner. Written by the renowned Professor Mohammad S. Obaidat and his coauthor Professor Noureddine Boudriga, it is also the only resource to treat computer and telecommunication systems as inseparable issues. The authors explain the basic concepts of performance evaluation, applications, performance evaluation metrics, workload types, benchmarking, and characterization of workload. This is followed by a review of the basics of probability theory, and then, the main techniques for performance evaluation namely measurement, simulation, and analytic modeling with case studies and examples. Contains the practical and applicable knowledge necessary for a successful performance evaluation in a balanced approach Reviews measurement tools, benchmark programs, design of experiments, traffic models, basics of queueing theory, and operational and mean value analysis Covers the techniques for validation and verification of simulation as well as random number generation, random variate generation, and testing with examples Features numerous examples and case studies, as well as exercises and problems for use as homework or programming assignments Fundamentals of Performance Evaluation of Computer and Telecommunication Systems is an ideal textbook for graduate students in computer science, electrical engineering, computer engineering, and information sciences, technology, and systems. It is also an excellent reference for practicing engineers and scientists.
This Festschrift volume is published in honor of Günter Haring on the occasion of his emerital celebration and contains invited papers by key researchers in the field of performance evaluation presented at the workshop Performance Evaluation of Computer and Communication Systems - Milestones and Future Challenges, PERFORM 2010, held in Vienna, Austria, in October 2010. Günter Haring has dedicated most of his scientific professional life to performance evaluation and the design of distributed systems, contributing in particular to the field of workload characterization. In addition to his own contributions and leadership in international research projects, he is and has been an excellent mentor of young researchers demonstrated by their own brilliant scientific careers. The 20 thoroughly refereed papers range from visionary to in-depth research papers and are organized in the following topical sections: milestones and evolutions; trends: green ICT and virtual machines; modeling; mobility and mobile networks; communication and computer networks; and load balancing, analysis, and management.
"Reliable performance evaluations require the use of representative workloads. This is no easy task since modern computer systems and their workloads are complex, with many interrelated attributes and complicated structures. Experts often use sophisticated mathematics to analyze and describe workload models, making these models difficult for practitioners to grasp. This book aims to close this gap by emphasizing the intuition and the reasoning behind the definitions and derivations related to the workload models. It provides numerous examples from real production systems, with hundreds of graphs. Using this book, readers will be able to analyze collected workload data and clean it if necessary, derive statistical models that include skewed marginal distributions and correlations, and consider the need for generative models and feedback from the system. The descriptive statistics techniques covered are also useful for other domains"--
An overview of queueing network modelling. Conducting a modelling study. Fundamental laws. General analytic technique. Bounds on performance. Models with one job class. Models with multiple job classes. Flow equivalence and hierarchical modelling. Representing specific subsystems. Memory. Disk I/O. Processors. Parameterization. Existing systems. Evolving systems. Proposed systems. Perspective. Using queueing network modelling software. Appendices. Constructing a model from RMF data. An implementation of single class, exact MVA. An implementation of multiple class, exact MVA. Load dependent service centers. Index.
This book constitutes the thoroughly refereed post-conference proceedings of the 24th International Workshop on Job Scheduling Strategies for Parallel Processing, JSSPP 2021, held as a virtual event in May 2021 (due to the Covid-19 pandemic). The 10 revised full papers presented were carefully reviewed and selected from 17 submissions. In addition to this, one keynote paper was included in the workshop. The volume contains two sections: Open Scheduling Problems and Proposals and Technical Papers. The papers cover such topics as parallel computing, distributed systems, workload modeling, performance optimization, and others.