Thesis (M.S.)--University of Hawaii at Manoa, 2006. / Includes bibliographical references (leaves 60-62). / vii, 102 leaves, bound ill. 29 cm

published_or_final_version / Electrical and Electronic Engineering / Doctoral / Doctor of Philosophy

A global schema is an integrated view of heterogeneous databases used to support data sharing among independent, existing databases. Global schema design complexities arise from the volume of details, design choices, potential conflicts, and interdependencies among design choices. The methodology described provides a framework for efficient management of these critical dimensions in generating and evaluating alternative designs. The methodology contains three major steps. First, differences due to the varying local data models are resolved by converting each local schema to an equivalent schema in a unifying data model. Second, the entity types of the local schemas in the unifying model are grouped into clusters called common areas. All the entity types in a common area can possibly be merged via generalization. For each common area, semantic information is defined that drives the merging process. Third, each common area is integrated into the global schema by applying a set of generalization operators. Mapping rules are then defined to resolve differences in the representations of equivalent attributes. Th integration of the local schemas is based on equivalence assertions. Four types of attribute equivalences are defined: two attributes may be locally or globally equivalent, and they can be key or non-key. Strategies for handling each of these cases are proposed and evaluated. The global schema design methodology includes several algorithms which may assist a designer. One algorithm analyzes a set of equivalence assertions for consistency and completeness including resolution of transitively implied assertions. A second algorithm performs an interactive merge of a common area by presenting the possible generalization actions to the designer. It supports the theme that many generalization structures can be possible, and the appropriate structure often depends on designer preferences and application requirements. The methodology is evaluated for several cases involving real databases. The cases demonstrate the utility of the methodology in managing the details, considering many alternatives, and resolving conflicts. In addition, these cases demonstrate the need for a set of computer-aided tools; for even a relatively small case, the number of details and design choices can overwhelm a designer.

A high speed, 125 mega-bit per second data rate, data communication channel using fiber optic technology is described. Medical image data, generated by CT scanner or magnetic resonance imaging type imaging equipment, passes from standard American College of Radiology - National Electrical Manufactures Association (ACR-NEMA) interface equipment to the High Speed Fiber Optic Network Interface (HSFONI). The HSFONI implements the ACR-NEMA standard interface physical layer with fiber optics. The HSFONI accepts data from up to 8 devices and passes data to other devices or to a data base archive system for storage and future viewing and analysis. The fiber components, system level, and functional level considerations, and hardware circuit implementation are discussed.

This paper describes the design and simulation of a network interface unit (NIU) for a picture archiving and communication system (PACS) network called PACnet. PACnet is a dual fiber optic ring network under development at the Computer Engineering Research Laboratory of the University of Arizona. This network integrates voice, data, and image communications in a hospital environment and supports a throughput rate between 200-500 megabits per second. At each node in the network, an NIU implements the Data Link Layer and Physical Layer protocols of PACnet. The initial network interface unit design for PACnet was a functional description of NIU protocols and major components. In order to construct a demonstration prototype of PACnet,the NIU description must be refined and an architecture must be specified. The NIU design is specified and simulated using the hardware description language VHDL. Simulation results provide information on NIU timing characteristics and logic families required to implement the NIU.

This thesis deals with the analysis of a Wired High Speed Serial Data Link (PAM2) which is commonly used throughout the data-communications and tele-communications industry. The goal of this study is to build a scalable simulation tool using Matlab that ultimately uses Receiver Bit Error Ratio (BER) as the metric for data link health. This study is also designed to aid in link specification development. The Matlab and theoretical development is broken up into three sections being Transmitter (TX), Channel (Hs) and Receiver (RX). Realistic noise impairments can be added to each section along the signal path creating signal stresses commonly seen in data center applications. The TX function is designed to create random and periodic timing jitter, voltage noise and deterministic pre-distortion filtering effects. For the channel response s-parameters are used as the model result for many commonly seen channel loss and reflection scenarios. The RX model uses signal to noise ratio and vertical eye margin to determine the equalized link BER. The study results show many tradeoffs between noises, RX Equalizer, RX gain and RX BER. The simulation results also reveal that there is no closed form solution for converging the modern closed-eye PAM2 detector.

We have investigated, in various multiple senses, the “capacity” of several models of erasure networks. The defining characteristic of a memoryless erasure network is that each channel between any two nodes is an independent erasure channel. The models that we explore differ in the absence or presence of interference at either the transmitters, the receivers, or both; and in the availability of feedback at the transmitters. The crux of this work involves the investigation and analysis of several different performance measures for these networks: traditional information capacity (including multicast capacity and feeback capacity), secrecy capacity, and transport capacity. / text