Modeling Techniques and Tools for Computer Performance Evaluation

fhere is a growing perception that the succesful design and ~xploitation of complex information systems is critically dependant on :he level of methods and support for performance modelling. Fast :echnological evolution creates a strong demand for analysis :echniques which can handle new hardware and software )rganisations. Performance modelling users are expecting improved :ools in terms of their user friendliness and their integration in the ;ystem design and operation process. Providing an image as clear as possible of the recent advances in :hese areas is the objective of the serie of In t ern a t ion a 1 Conference on Modelling Techniques and Tools for Computer Performance Evaluation. We believe thst the 4th edition does respond to the objective. It illustrates the most recent trends in our jomain and gives insight into what we may expect in terms of new modelling techniques; tools and practices in the future years. The proceedings include the twenty seven papers accepted by the Program Committee (but of sixty six submitted) and three invited papers presented at the Conference. They cover the following topics: Modelling Tools Networking Architecture and Workload Protocols Architecture Configuration Parallelism ix Performability Algorithms We would like to express our sincere apprecIatIOn and gratitude to the members of the Program Committee and to the reviewers for their time and effort. Without their kind cooperation, the Conference would not have been possible.

Inhalt
Modelling Tools.- The Challenge of Combining Text and Graphics in Modelling Support Environment.- Compiling Techniques for the Analysis of Stochastic Petri Nets.- PANACEA: An Integrated Set of Tools for Performance Analysis.- Distributed Systems Design: an Integrated Software Tool for Performance Analysis.- Towards a Performance Modelling Environment: News on HIT.- Networking.- LANSF A Modular Approach for Modeling Low-level Communication Protocols.- IPwatch: A Tool for Monitoring Network Locality.- QUARTZ: A Tool for Performance Evaluation of Data Networks.- Architecture and Workload.- TEL: A versatile Tool for Emulating system Load.- A Tool for Performance-Driven Design of Parallel Systems.- Protocols.- Performance Analysis of a Buffer under Locking Protocols.- Average Message Delay of an Asymmetric Single-Buffer Polling System with Round-Robin Scheduling of Services.- On a Multi-Class Queue with Class Dependent Window-Flow Control.- Architecture Configuration.- Solving some Practical Problems of Computing System Configuration and Design by Throughput Control in Closed Separable Queueing Networks.- Modeling DASD Configurations with Multipathing and Multiple Blocking Points.- A Unified Approach to the Performace Evaluation of Multiprocessors.- A Hardware Instrumentation Approach for Performance Measurement of a Shared-Memory Multiprocessor.- Parallelism.- Artificial Intelligence Modelling of Complex Systems.- PEPS: A Package for Solving Complex Markov Models of Parallel Systems.- Performance of Concurrent Redezvous Systems with Complex Performance Structures.- On the Tradeoff between Parallelism and Communication.- Performability.- PROPER, A Performability Modelling and Analysis Tool.- Petri Nets Generating Markov Reward Models for Performance/ReliabilityAnalysis of Degradable Systems.- Failure Data Analysis of a Computer System.- Algorithms.- Fast Approximate Solution of Queueing Networks with Multi-Server Chain-Dependent FCFS Queues.- A Universal Maximum Entropy Algorithm for General Multiple Class Open Networks with Mixed Service Disciplines.- Piecewise Approximation of Densities Applying the Principle of Maximum Entropy: Waiting Times in G/G/1-Systems.- Two-Stage Cyclic Network with Blocking: Cycle Time Distribution and Equivalent Properties.