In the new and sophisticated cyber attacks (mainly, advanced persistent threats) the advanced specimens of malware such that zero-day malware play a crucial role. Due to its stealthy behavior it is very important to study and analyze its propagation process by designing mathematical models that could predict in an efficient way its spread on a network. With no doubt the computational implementation of these theoretical models leads to the develop of solutions to be used in the Security Operation Centers (SOC) with forensic purposes. The main goal of this work is to introduce a novel mathematical model to simulate advanced malware. Specifically, it is a compartmental and global SCIRAS (Susceptible-Carrier-Infectious-Recovered-Attacked-Susceptible) model where susceptible, carrier, infectious, recovered and attacked devices are considered. The local and global stability of its equilibrium points are studied and the basic reproductive number is computed. From the analysis of this epidemiological threshold, the most efficient security countermeasures are derived.