Abstract
Organisations have become more reliant on electronic assets in recent years, as a shift in focus has driven organisations to make extensive use of Critical Information Infrastructure (CII) to drive various business activities. While there has been a significant paradigm shift during this transition, most organisations have failed to ensure that sufficient security mechanisms are put in place to protect the organisation and their CII from exploitation.
Typically, these organisations employ conventional security mechanisms such as a firewall, proxy or anti-virus software, but these approaches are fallible. An organisation can simply not afford to have its CII exploited, as this results in reputational and financial losses. Every single organisation should define their appetite for risk by performing a Risk Value Assessment. (RVA). Unfortunately, it is impossible for an organisation to protect its CII against every possible threat, as threats are polymorphic and dynamic in nature.
The research proposes a hybrid approach towards improving the Critical Information Infrastructure Protection (CIIP) capabilities within an organisation. The Continual Evaluative Self-aware Immune-inspired Multi Agent Critical Information Infrastructure Protection System (CESIMAS) utilises various concepts and ideal analogies from the research fields of Multi Agent Systems, Artificial Immune Systems, Self-awareness, and Ambient Intelligence to propose a hybrid virtualised metaphysical model. The CESIMAS model utilises various sub-systems and agent types to establish an automated, self-sufficient and self-regulatory eco-system whereby the agents in the model effectively and efficiently attempt to provide an improved CIIP capability within an organisation’s Critical Information Infrastructure.
The CESIMAS model contributes a virtualised meta-physical model, which illustrates how an Ambient Intelligence-based approach can be implemented and modelled to potentially improve the level of CIIP within an organisation.
The CESIMAS model proposes and contributes a more efficient and effective agent generation process, parts of which are utilised to improve immune-inspired detection techniques within the model. The model establishes a hybrid approach to self-set maintenance and immune-inspired detection techniques, whilst reducing the computational penalties and constraints.
Ph.D. (Computer Science)