Please use this identifier to cite or link to this item:
Title: Task Oriented Fault-Tolerant Distributed Computing for Use on Board Spacecraft
Authors: Fayyaz, Muhammad
Supervisors: Vladimirova, Tanya
Warrington, Michael
Award date: 4-Jan-2016
Presented at: University of Leicester
Abstract: Current and future space missions demand highly reliable, High Performance Embedded Computing (HPEC). The review of the literature has shown that no single solution could meet both issues efficiently at present addressing HPEC as well as reliability. Furthermore, there is no suitable method of assessing performance for such a scheme. In this thesis a novel cooperative task-oriented fault-tolerant distributed computing (FTDC) architecture is proposed, which caters for high performance and reliability in systems on board spacecraft. In a nut shell, the architecture comprises two types of nodes, a computing node and an input-output node, interfaced together through a high-speed network with bus topology. To detect faults in the nodes, a fault management scheme specifically designed to support the cooperative task-oriented distributed computing concept is proposed and employed, which is referred to as Adaptive Middleware for Fault-Tolerance (AMFT). AMFT is implemented as a separate hardware block and operates in parallel with the processing unit within the computing node. A set of metrics is designed and mathematical models of availability and reliability are developed, which are used to evaluate the proposed distributed computing architecture and fault management scheme. As a new development, extending the current state of the art, the proposed fault-tolerant distributed architecture has been subjected to a rigorous assessment through hardware implementation. Implementation approaches at two levels were adopted to provide a proof of concept: a board level and a Multiprocessor System-on-Chip (MPSoC) level. Both distributed computing system implementations were evaluated for functional validity and performance. To examine the FTDC architecture performance under a realistic space related distributed computing scenario a case-study application, representing a satellite Attitude and Orbit Control System (AOCS), was developed. The AOCS application was selected because it features a time critical task execution, in which system failure and reconfiguration time must be kept minimal. Based on the case-study application, it was demonstrated that the FTDC architecture is capable of fully meeting the desired requirements by timely migrating tasks to functional nodes and keeping rollback of task states minimal, which proves the advantages of the adopted cooperative distributed approach for use on board spacecraft.
Type: Thesis
Level: Doctoral
Qualification: PhD
Rights: Copyright © the author. All rights reserved.
Appears in Collections:Theses, Dept. of Engineering
Leicester Theses

Files in This Item:
File Description SizeFormat 
2015fayyazmfphd.pdf5.52 MBAdobe PDFView/Open

Items in LRA are protected by copyright, with all rights reserved, unless otherwise indicated.