Intelligent Active Queue Management Using Explicit Congestion Notification

As more end devices are getting connected, the Internet will become more congested. A variety of congestion control techniques have been developed either on transport or network layers. Active Queue Management (AQM) is a paradigm that aims at mitigating the congestion on the network layer by active buffer control to avoid overflow. However, finding the right parameters for an AQM scheme is challenging, due to the complexity and dynamics of the networks. On the other hand, the Explicit Congestion Notification (ECN) mechanism is a solution that makes visible incipient congestion on the network layer to the transport layer. In this work, we propose to exploit the ECN information to improve AQM algorithms by applying Machine Learning techniques. Our intelligent method uses an artificial neural network to predict congestion and an AQM parameter tuner based on reinforcement learning. The evaluation results show that our solution can enhance the performance of deployed AQM, using the existing TCP congestion control mechanisms.

Optimal Policy of Window Flow Control Based on Packet Transmission Interval with Explicit Congestion Notification

This paper discusses a reliability problem of window flow control scheme with Explicit Congestion Notification (ECN) in the communication system, which is an essential issue in the long-distance transmission such as cloud computing. For example, there is a problem of reduction in transmission efficiency caused by a packet loss from network congestion. In order to solve the problem, recently, High-performance and Flexible Protocol (HpFP) has been proposed and performed a verification of effectiveness by simulation. 6 HpFP provides efficient packet communication by changing a packet transmission interval [T. Murata, T. Mizuhara, A. Takagi, K. Fukushima, K. Yamamoto, Y. Nagaya, K. Muranaga and E. Kimura, HpFP: A new protocol for LFNs with packet-loss based on UDP: A basic concept and detailed design of the protocol, IEICE Technical Report 115(484) IN2015-109, pp. 7–12 (2016)]. We have discussed the reliability model of a window flow control scheme based on a packet transmission interval. 7 We have proposed the model which judges the occurrence of congestion in a network by the number of retransmission [M. Kimura, M. Imaizumi and T. Nakagawa, Reliability modelling of window flow control scheme for a communication system based on packet transmission interval, in Proc. 23rd ISSAT Int. Conf. Reliability and Quality in Design (2017), pp. 171–175]. On the other hand, ECN mechanism for the detection of incipient congestion has been already proposed. ECN mechanism is efficient for the window flow control scheme because ECN bit is set in the packet header as an indication of congestion and is avoided unnecessary packet drops. We consider a stochastic model of a window flow control scheme based on a packet transmission interval with ECN. If ECN bit has been set during connection, a packet transmission interval is prolonged. We derive the mean time until packet transmissions succeed, and analytically discuss an optimal policy to maximize the amount of packets per unit of time until the transmission succeeds.