Several access control models have been developed: discretionary, mandatory, role-based, organisation-based or the recent hierarchical organisation-based. The hierarchical organisation-based model is best suited to counter internal attacks and fraud because it integrates a digital parafer that performs proactive hierarchical control of all activities carried out by the organisation's employees in the system, including those of the super-user. However, because this control is based solely on the type of requests issued by internal employees, this model is vulnerable to Denial of Service attacks in a Big Data environment. The purpose of this article is to optimise the number of requests awaiting hierarchical treatment in the hierarchical organisation-based access control model while maintaining a maximum level of security. For this purpose, we have proposed an additional layer of control of the digital parafer based on suspicious behaviour detection from the entropy measurement. Finally, we carried out a mathematical and experimental comparisons between the two approaches in terms of number of requests awaiting treatment in the system and obtained a satisfactory result.