published_or_final_version / Mechanical Engineering / Master / Master of Philosophy

This dissertation describes the design and implementation of the file system component of the Saguaro operating system for computers connected by a local-area network. Systems constructed on such an architecture have the potential advantage of increased file availability due to their inherent redundancy. In Saguaro, this advantage is made available through two mechanisms that support semi-automatic file replication and access: reproduction sets and metafiles. A reproduction set is a collection of files that the system attempts to keep identical on a "best effort" basis, relying on the user to handle unusual situations that may arise. A metafile is a special file that contains symbolic path names of other files; when a metafile is opened, the system selects an available constituent file and opens it instead. These mechanisms are especially appropriate for situations that do not require guaranteed consistency or a large number of copies. Other interesting aspects of the Saguaro file system design are also described. The logical file system forms a single tree, yet any file can be placed in any of the physical file systems. This organization allows the creation of a logical association among files that is quite different from their physical association. In addition, the broken path algorithm is described. This algorithm makes it possible to bypass elements in a path name that are on inaccessible physical file systems. Thus, any accessible file can be made available, regardless of the availability of directories in its path. Details are provided on the implementation of the Saguaro file system. The servers of which the system is composed are described individually and a comprehensive operational example is supplied to illustrate their interation. The underlying data structures of the file system are presented. The virtual roots, which contain information used by the broken path algorithm, are the most novel of these. Finally, an implementation of reproduction sets and metafiles for interconnected networks running Berkeley UNIX is described. This implementation demonstrates the broad applicability of these mechanisms. It also provides insight into the way in which mechanisms to facilitate user controlled replication of files can be inexpensively added to existing file systems. Performance measurements for this implementation are also presented.

This dissertation is concerned with the design of a high speed fiber optic campus backbone network which not only provides a high bandwidth connection to various LANs, PBXs, ISDNs, supercomputers, and other high speed computing facilities but also offers an integrated service of voice and data. The campus backbone network is based on an optical fiber dual ring structure. A new token ring protocol based on the IEEE 802.5 standard is proposed as the medium access protocol for the backbone network. In the proposed token ring network, the token holder transmits packets in an exhaustive way when no other nodes want to transmit voice packets. If any node has voice packets waiting for transmission, the token holder can transmit only a single packet during its access opportunity. In this way, not only can the voice delay be bounded at high voice traffic intensities, but a high data throughput can be achieved at low voice traffic intensities. The potential use of Time Division Multiplexing (TDM) was also considered, but the TDM technique was found to be less suitable for our campus environment mainly due to its low efficiency in bursty environments and complex hardware requirements. The backbone network was simulated using the Simscript II.5 discrete event simulation language. The simulation results indicated that the proposed token ring protocol provides satisfactory performances in every respect. A network interface unit (NIU) was designed, which consists of optical transmitter, receiver, and fault tolerance mechanisms as well as the protocol engine. The optimal values of major parameters for the NIU design were determined by the simulation.

Adaptive routing and flow control methods are necessary in computer networks due to the nonstationary conditions that exist in such networks. In this dissertation three distinct yet complementary approaches to modeling computer networks for performance evaluation and control under nonstationary conditions are presented namely: queueing theory, discrete event simulation, and state variable modeling. The application of these techniques to the design and performance evaluation of adaptive routing and flow control methods is demonstrated with the development of a new two-level hierarchical adaptive buffer management scheme and a dynamic virtual circuit routing policy.

This research identifies a current and future need in the realm of information systems development which has surfaced as a result of layered architectures and software reuse. An analysis methodology based upon two three-dimensional metamodels which correspond to the principal aspects of system architecture, structure and communication, is developed. Each metamodel can be viewed as having three planes which represent increasing abstractions away from actual source code. For example, with regard to the structure metamodel, the lowest plane corresponds to actual source code structures written in a specific computer language, the middle plane represents the general form of the structure available in that language, and the top plane represents the general form of structures available in any language. An object-oriented viewpoint was adopted in order to allow the expression of the relationships between entities found on a single plane of a metamodel, as well as the expression of the relationships between entities found on different planes. The metamodels provide a framework and methodology for discerning the structure and communication mechanisms employed in software source code as well as a framework from within which behavioral models can be developed.

Recent developments in information systems technology have made it possible for individuals to work together anonymously using networked personal computers. In this dissertation, a theory of anonymous interaction is proposed. Evidence is provided to suggest that anonymity has deindividuating effects on group process and can, therefore, influence group outcomes in several ways. Two experiments on anonymity in idea generation are presented. In Study 1, where subjects could leave at their discretion, identification kept them longer and caused them to type more, though there were no differences in the quantity or quality of the ideas across experimental conditions. In Study 2, where subjects were forced to stay, identifiability lost importance. Responsibility, however, rose in importance. Subjects with sole responsibility for their task produced more output than did subjects who shared responsibility. Taken together, these results forced us to reject the hypothesis that anonymous subjects would produce more output than would identified subjects. These results show that we cannot speak simply of the effects of anonymity on idea generation in computer-supported groups. With a straightforward interpretation of previous experiments on GDSS anonymity, it was hypothesized that anonymous subjects would produce more than identified subjects. They did not. It is clear that anonymity will lead to deindividuation, enabling participants to engage in uninhibited behavior. However, whether their behavior is productive or unproductive is determined, at least in part, by task, interaction, and technology.

This dissertation describes queueing-theoretic analysis of shared service systems that require that customers leave the system in the sequence in which they arrived. This requirement makes it necessary to resequence customers before they leave the system. Resequencing adds new complications to the analysis of queueing systems. While waiting time is still important, resequencing results in a new type of "non-working" delay of a customer called the resequencing delay. This dissertation presents primarily analytical and numerical methods to determine the distribution and mean value of resequencing delay, and of total delay. In the simplest models closed form analytical expressions have been obtained, but in more complex models numerical methods have been developed to compute the distribution and mean of resequencing delay, and of total delay. This enables us to study the behavior of resequencing and total delay as system parameters are changed. For several composite server models we present expressions for the distribution and mean of resequencing delay, and of total delay. In particular we consider the M/M/∞ composite server model, the M/H(K)/∞ composite server model, the G/M/∞ composite server model, the M/M/m composite server model, and the G/M/m composite server model. The formulas are interpreted using asymptotic approximation or bounding techniques. For more general composite server models, it is difficult to obtain closed form expressions for resequencing and total delay. We develop numerical methods based on matrix-geometric methods to compute resequencing and total delay. In particular, we develop numerical methods for the computation of the mean resequencing delay, and mean total delay for the M/H₂/m composite server model, and the M/Hypo₂/m composite server model.

Information system requirements determination is a key area in management information systems research that includes the problems of requirements specification, requirements elicitation, and user involvement. The combination of these three problems is a research area which we call the group approach to information system requirements determination. The main contribution of this research is a model to be used for the problem of fitting existing computer aids to this research area and a set of methods for solving the usability and configuration problems when using such computer aids. The usability problem is that of determining whether a set of computer aids can be used effectively in accomplishing the task of requirements determination. The configuration problem is that of selecting a minimum collection of funtionalities necessary for economically supporting all aspects of requirements determination. Electronic meeting systems are the general category of computer aids that we are interested in applying to the task. In particular, the GroupSystems electronic meeting system developed at The University of Arizona is used as a case in this research. Characteristics of the requirements determination task and profiles of GroupSystems tools are combined into our model for fitting computer aids to a given task. We then derive the answers regarding the usability and configuration of GroupSystems in the group approach to information system requirements determination. We also compare the derived configurations to the GroupSystems tools used in an authentic case. The main points of this research include: (1) proposal of the concept of fitting computer aids to, instead of developing new computer aids for, a task area; (2) proposal of a model for fitting computer aids via a classification of information types; (3) analysis of the natures of the fitting model and the usability and configuration problems; (4) proposal of approximation methods for solving the configuration problem; (5) analysis of the task area--the group approach to information system requirements determination; (6) analysis of the use of GroupSystems in the task area by applying the proposed fitting model and approximation methods; and (7) demonstration of the usefulness of the fitting model and approximation methods by analyzing an authentic case of using GroupSystems tools.

This dissertation describes LABNET, a loosely-coupled network of small computers for laboratory automation and data collection. The network comprises two parts: RAPNET, the local-network operating-system-like software, and Real-time MICRODARE, an interactive language for programming automation and data-collection tasks. RAPNET provides the framework upon which application-level programs like MICRODARE execute. In addition to the usual file services and other miscellaneous system services normally supplied by a single-CPU operating system, RAPNET provides link-level message facilities, program control, and a virtual channel system. There is a means for coordinated application-level program intercommunication, the pseudo-link; pseudo-links are the means by which programs running in different CPUs or in the same CPU may be connected. To the application-level program, a pseudo-link looks just like a file or device. Real-time MICRODARE supplies an interactive programming capability which uses the facilities of RAPNET to enable a programmer to do distributed program systems for automation, simulation, and data collection. MICRODARE consists of an interactive BASIC-like job-control language, and a compiled fast-task language. The job-control language permits time and event dependent scheduling of automation and data-collection program segments. The fast-task language does simulation, signal-processing, data-collection, and control tasks at close-to-assembly-language speeds.

No description available.