Degree Type

Dissertation

Date of Award

2011

Degree Name

Doctor of Philosophy

Department

Electrical and Computer Engineering

First Advisor

Diane T. Rover

Abstract

Today's embedded system designers face the challenges of ever increasing complexity and shorter time-to-market deadlines. System-level methodologies emerge to meet these challenges. Refinement-based methodologies, such as the SpecC methodology and Transaction Level Modeling, continue to gain popularity in the embedded system designers' community. However, as more communication-dominated applications and architectures appear in the market, designers find that the lack of models allowing system-level communication analysis is a major limiting factor in current system-level design methodologies. Thus, modeling for system-level communication analysis is key for a design methodology to thrive with today's embedded system designers. This work presents a new approach to system-level modeling that allows better communication analysis earlier in the design process. This approach defines a new model that utilizes random variables to include the communication details at higher abstraction levels. This work proposes a probabilistic model to include and evaluate the system communication features in the higher abstraction level. Guidelines to include the proposed model into a refinement-based methodology are presented, and methods for performance estimation are shown.

Copyright Owner

Ramon Alejandro Mercado

Language

en

Date Available

2012-04-28

File Format

application/pdf

File Size

96 pages

Share

COinS