Determining the Feasibility of Achieving Graceful System Degradation Through The Use of Automatic Reconfigurable Designs and Software

Cornwall, Mark D. (2006). Determining the Feasibility of Achieving Graceful System Degradation Through The Use of Automatic Reconfigurable Designs and Software. Student dissertation for The Open University module M801 MSc in Software Development Research Dissertation.

Please note that this student dissertation is made available in the format that it was submitted for examination, thus the author has not been able to correct errors and/or departures from academic standards in areas such as referencing.

DOI: https://doi.org/10.21954/ou.ro.0001603b

Abstract

All systems can and do fail, and often a relatively ‘small’ failure has a disproportionate effect, as it prevents access to parts of a system that are operating correctly, or prevents their usage. Hence, the traditional method of maintaining the intended service, is to provide sufficient redundancy, in combination with an appropriate repair strategy; although this is inherently expensive and is not always feasible. However, design and economic pressures, together with engineering’s innate conservatism, effectively prevent this approach from being re-evaluated. The net result is, at best, a step drop in capability and at worse, a total loss of a system. Allowing a system to undergo ‘graceful’ service degradation, instead of a step drop, significantly improves its ‘resilience’; with a direct impact on both availability and cost. A resilient system is therefore capable of absorbing the impact of an interruption, disruption or loss; and will continue to provide a minimum ‘acceptable’ level of service. This thesis is a literature survey, and represents a synthesis of ideas from a number of sources to suggest an inherently resilient system. It presents a new way for looking at system failures that occur within a traditional computer-based architecture, by considering: · An architecture based upon the ‘fractal concept’, i.e. a number of identical components, capable of self-repair, self-configuration, self-optimisation, and selfprotection; which through mutual interaction, can provide high level services equivalent to that of a ‘normal’ computer; · An operating system, which utilises the principles of evolutionary strategies in the form of genetic algorithms, to prioritise and reconfigure the components; which form the pool of available resources for a given operating scenario and environmental circumstances. By combining these two concepts, to create a ‘heterogeneous adaptive system’, this thesis outlines how the proposed approach creates an emergent self-optimising capability; which uses temporal based structures for causal elimination and automatic ‘work around’. This creates an inherently resilient system, capable of automatically adapting to internal and external environmental changes i.e. failures, while maintaining its intended service provisions. This thesis consists of six sections: the introduction; research methods employed; current architecture resilience concepts and responses; the proposed fractal architecture; an examination of the proposed architecture’s ability to address resilience issues; an analysis of the method employed and the results identified; conclusions and future research.

Viewing alternatives

Download history

Metrics

Public Attention

Altmetrics from Altmetric

Number of Citations

Citations from Dimensions

Item Actions

Export

About

  • Item ORO ID
  • 90171
  • Item Type
  • Student Dissertation
  • ISSN
  • 1744-1986
  • Extra Information
  • Published as Department of Computing Technical Report 2006/28
  • Academic Unit or School
  • Other Departments > Other Departments
  • Copyright Holders
  • © 2006 Mark D. Cornwall
  • Depositing User
  • Legacy Research Papers

Recommendations