Performability estimation of token ring network

number: 
1623
إنجليزية
Degree: 
Author: 
Rasha Albert Elia
Supervisor: 
Dr. Abdul-Kareem A. Najem Al-Aloosy
year: 
2006

Abstract : The work reported in this thesis deals with the performability evaluation of the Token Ring Network. Performability evaluation arose from the need to determine overall quality of service by relating and quantifying aspects of what a specific system is and does (how well it performs) with respect to what the system is required to be and to do (how its functionality is affected by faults). Performability is a term used to describe performance and reliability. Performance is defined as _quality of service provided the system is correct_. Reliability is _the property of a system which allows reliance to be justifiably placed on the service it delivers_. However, the problem that arises from evaluating these items separately is that performance depends on reliability. Thus, it was obvious that combining performance and reliability was needed for a complete evaluation; the issue was how to combine them. The method that is used for performability evaluation of token ring network is based on analysis describing the action of the token ring network. It was assumed that the Mean Time Frame Transfer (MTFT) to be the main measure of the performability which is composed of the Mean Waiting Time (MWT) for a free token and the Mean Real Transfer Time (MRTT) of a data frame (together with the acknowledgement receipt), taking in consideration all possible malfunctions. In addition, the influence of the effective parameters [number of nodes (N), length of the link (L), channel bandwidth (b), exponential distribution ratio of the user access probability (ë) and ratio of malfunctions (km)] on MTFT and its components were investigated. As a general conclusion the present investigation provides a new method for evaluation of performability of the token ring network. It was found that the Mean Time Frame Transfer and its two main components vary linearly with number of nodes, length of the link, exponential distribution ratio of the user access probability and the ratio of malfunction and vary inversely with channel bandwidth.