Files in this item



application/pdfMin Young_Nam.pdf (4MB)
(no description provided)PDF


Title:A tool for model-based engineering
Author(s):Nam, Min Young
Director of Research:Sha, Lui R.
Doctoral Committee Chair(s):Sha, Lui R.
Doctoral Committee Member(s):Johnson, Ralph E.; Caccamo, Marco; Bradford, Richard M.
Department / Program:Computer Science
Discipline:Computer Science
Degree Granting Institution:University of Illinois at Urbana-Champaign
Subject(s):Architecture Analysis and Design Language (AADL)
Model-Based Engineering
System Design Tool
Real-Time Systems
Safety Critical Systems
Itegrated Modular Avionics
Resource Allocation
Bus Arhictecture Design
I/O Design
System Level Analysis
Abstract:In many hard real-time avionics systems, more and more features are being added to faster but cheaper hardware. Thus, hardware resources such as computation and network bandwidth are increasingly being shared by multiple applications, leading to rapid increases in the size and complexity of the overall system. In the design and development process, system level performance analysis should precede implementation in order to save cost and to optimize the system architecture during a relatively earlier development phase when such changes are less costly to make. To provide system level performance analysis, Model-Based Engineering (MBE) or virtual integration is becoming more popular in the avionics industry as well as other manufacturers of Cyber-Physical Systems (CPS), which also include the automotive and medical industries. Virtual integration helps system designers to test multiple options, upgrade existing systems with modifications and check safety-critical, schedulability, and performance requirements relatively quickly using models. With a) the advance in high performance processors, including multi-core processors, b) the trend of designing Integrated Modular Avionics (IMA) systems that allows multiple applications to coexist on a platform, and c) the requirement for I/O intensive applications for better service, building an elaborate I/O architecture can bring great benefits in the overall design. Much previous work does not model the I/O architecture with sufficient detail to provide valuable feedback in improving the design, or it requires too much detail that it cannot be used for early analysis. This dissertation describes the problems that we have been addressing in the current area of utilizing architectural Model-Based Engineering (MBE) for avionics systems and the solutions that we have developed along the way. First, we show a framework including our tool named ASIIST (Application Specific I/O Integration Support Tool) which is a MBE analysis tool that we have developed for I/O architecture design and integrated processor and I/O scheduling. The computer-aided I/O design framework demonstrates what elements need to be considered when creating a software analysis tool taking benefit of MBE. Later we present how ASIIST supports IMA systems in ways of handling side effects of I/O traffic that are known but difficult to keep up-to-date due to complexity. In addition, we provide a framework for auto-generation of a robust tree-shaped I/O architecture where the user can easily search for an architecture that can improve the tolerance against I/O workload uncertainties, especially for an IMA system. ASIIST allows for rapid evaluation of schedulability impacts of I/O architecture designs in the context of IMA. Here we do a design space search approach and investigate how MBE can be applied. This early consideration of I/O effects would increase the likelihood that the system remains schedulable after actual coding and integration of multiple applications. Finally, we have built a model-based approach to extend existing analysis algorithms. The extensibility of Architecture Analysis and Design Language (AADL) and the loose syntax rules can often lead to difficulties for actual users to make changes to a existing model without the worry of making existing analysis tools break. Thus, a concept of Open Analytic Runtime (OAR) Models is created where we express the details of analysis algorithms in the modeling components. This provides an open development environment where anybody can inspect and update the analysis implementation from the model. The analysis annex, is an implementation of OAR models for AADL. With the capability of modeling analysis algorithms, we were able to provide services that assist in the implementation of multiple correlated analysis algorithms. The analysis annex solver can search for correct order of execution for analysis algorithms and also verify the compatibility of assumptions from Resource Allocation (RA) algorithms and also connect assumptions of algorithms with the properties (guarantees) that other algorithms provide in order to ensure that all assumptions and all runtime quality attributes (RQA) are satisfied.
Issue Date:2013-02-03
Rights Information:Copyright 2012 Min Young Nam; ASIIST: Application Specific I/O Integration Support Tool for Real-Time Bus Architecture Designs, Proceedings of the 14th IEEE International Conference on Engineering of Complex Computer Systems, pages. 11-22, 2009 IEEE. Reprinted by permission.; In reference to IEEE copyrighted material which is used with permission in this thesis, the IEEE does not endorse any of University of Illinois at Urbana Champaign's products or services. Internal or personal use of this material is permitted. If interested in reprinting/republishing IEEE copyrighted material for advertising or promotional purposes or for creating new collective works for resale or redistribution, please go to to learn how to obtain a License from RightsLink.; Resource allocation contracts for open analytic runtime models. In Proceedings of the 9th ACM International Conference on Embedded Software (EMSOFT '11). pages 13-22. 2011 ACM, Inc. Reprinted by permission.;
Date Available in IDEALS:2013-02-03
Date Deposited:2012-12

This item appears in the following Collection(s)

Item Statistics