Browse Prior Art Database

PERFORMABILITY MODELING OF DISTRIBUTED REAL-TIME SYSTEMS

IP.com Disclosure Number: IPCOM000128457D
Original Publication Date: 1983-Aug-01
Included in the Prior Art Database: 2005-Sep-16
Document File: 14 page(s) / 64K

Publishing Venue

Software Patent Institute

Related People

Meyer, J.F.: AUTHOR [+3]

Abstract

This survey/position paper concerns modeling concepts and techniques for the performability (performance-reliability) evaluation of distributed real-time systems. Due to the nature of application requirements, such systems typically exhibit properties of concurrency, timeliness, fault tolerance, and degradable performance. Relevant prior research is surveyed and classified according to these properties and, based on the position that performability models should accommodate all four properties, directions for future research are suggested.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 7% of the total text.

Page 1 of 14

THIS DOCUMENT IS AN APPROXIMATE REPRESENTATION OF THE ORIGINAL.

PERFORMABILITY MODELING OF DISTRIBUTED REAL-TIME SYSTEMS

J. F. Meyer

CRL-TR-28-83

AUGUST 1983

THE UNIVERSITY OF MICHIGAN COMPUTING RESEARCH LABORATORY Room 1079, East Engineering Building
Ann Arbor, Michigan 48109
USA
Tel: (313) 763-8000

Invited Paper: International Workshop on Applied Mathematics and Performance/Reliability Models of Computer/Communication Systems
Pisa, Italy
Sept. 1983

ABSTRACT

This survey/position paper concerns modeling concepts and techniques for the performability (performance-reliability) evaluation of distributed real-time systems. Due to the nature of application requirements, such systems typically exhibit properties of concurrency, timeliness, fault tolerance, and degradable performance. Relevant prior research is surveyed and classified according to these properties and, based on the position that performability models should accommodate all four properties, directions for future research are suggested.

1. INTRODUCTION

The inherent complexity of distributed real-time systems, coupled with rapidly increasing demands for their application, have caused their investigation to become a subject of recognized importance. In this paper, we consider problems of modeling such systems for the purpose of evaluating their ability to perform, i.e., their performability (a unified performance- reliability concept which includes, as special cases, the usual notions of performance and reliability). More specifically, we are concerned with the development of mathematical models and model-based methodologies for evaluating the performability of distributed systems designed for real-time applications, e.g., computer-communication networks, vehicle control systems, industrial process-control systems, automated manufacturing systems, etc.

The intent of this paper is to (i) identify properties of such systems that need to be accounted for in the modeling process, (ii) survey and classify prior research according to the identified properties, and (iii) suggest directions for future research based on existing model deficiencies.

University of Michigan Computing Research Laboratory Page 1 Aug 01, 1983

Page 2 of 14

PERFORMABILITY MODELING OF DISTRIBUTED REAL-TIME SYSTEMS

II. SYSTEM PROPERTIES

Performance-reliability requirements for distributed real-time systems, although differing in detail from application to application, have some typical properties which, in turn, impose special conditions on system structure and behavior. in particular, such requirements typically call for high performance when the system is fault-free. This, coupled with the fact that resources are distributed, results in systems which exploit concurrent (parallel) processing. Secondly, real-time aspects of such requirements often impose conditions on the "timeliness" of processing activities, e.g., requirements for maximum allowable (input-output) response times which impose deadlines on the...