Anomaly Detection for Soft Security in Cloud Based Auditing of Accounting Systems

Achieving information security in the cloud is not a trivial exercise. When the systems involved are accounting software systems, this becomes much more challenging in the cloud, due both to the systems architecture in use, the challenges of proper configuration, and to the multiplicity of attacks that can be made against such systems. A particular issue for accounting systems concerns maintaining a proper audit trail in order that an adequate level of audit may be carried out on the accounting records contained in the system. In this paper we discuss the implications of the traditional approach to such systems and propose a complementary soft security solution relying on detecting behavioural anomalies by evidence theory. The contribution is in conceptualising the anomalies and providing a somewhat theoretical solution for a difficult and challenging problem. The proposed solution is applicable within any domain consisting of rigorous processes and risk of tampering or data exfiltration, such as the cloud based accounting systems.