Uneven growth rates of data links will inevitably leads to a "narrow" places in the info-communications network and thus to congestion, especially when connecting network access layer to the transport network. Traditional protocols queue management and congestion can't cope with the traffic management of high density and non-linear load, which can lead to congestion that causes a decrease in the effective data rate and degrades quality parameters, such as the percentage of packet loss, delay, and delay variation. Most often in data networks is using a passive mechanism of droping the packets when the queue is become full. When sending TCP traffic, this mechanism leads to the appearance the phenomenon of global synchronization. In order to early prevention of queue overflow, have been developed mechanisms for early detection congestion. This article describes the principles of building a simulation model for active queue management algorithms. The algorithms and program listings performing discrete-event simulation. These algorithms are used to perform simulation using high-level languages. Increasing complexity of modern telecommunication technologies creates a problem in the application of analytical techniques to evaluate the performance of designed systems and networks, and causes extensive use of simulation. In this paper in order to control congestion the nodes packet-switched networks using active queue management algorithms. The relevant service system, which simulates the work of a network node, each incoming packet can be dropped with a certain probability, even if the buffer is not yet fully filled. Probability of dropping depends on the length of the queue at the time of income the package. The paper presents the principles of simulation of discrete-event simulation algorithms of active queue management. The algorithms which perform simulation algorithms for highlevel languages are considered.
