MULTI-AGENT BASED INTRUSION DETECTION SYSTEM FOR CLOUD COMPUTING

Abstract

Cloud Computing has come to represent a significant shift in Information Technology. Companies are excited bythe opportunities to reduce capital costs, invest on core competencies and as well by the agility offered by therapid on-demand provisioning of computing resources. However, security concerns have slowed down the uptakeand trust has been identified as the key barrier to the uptake by most small to medium sized enterprises. Thispaper details a multi agent based framework for intrusion detection in the cloud computing environment thatprovides a mechanism for auditability and accountability. The design and development of the framework is basedon PASSI (Process for Agent Societies Specification and Implementation) methodology. The proposed designconsists of a virtual environment used to simulate the cloud computing environment, client and server side agent(java based) systems setup in the virtual instances. The outcome of the experiments carried out reveal thatmonitoring of all user operations on files predefined for monitoring within the virtual instances was successful.Email and SMS notifications were sent in real time to either clients or administrators for any operations detected.The framework depicts aspects of security assurance that assists in the efforts to promote trust between providersand consumers of cloud computing hence increase the uptake of the technology. The potential beneficiaries of theimplementation of the framework are both service providers and consumers of cloud computing.