For the widely deployed on-chip shared buffer, efficient buffer management is the key to absorbing bursts and avoiding packet loss during transient congestion. However, as the buffer-per-port-per-Gbps in production data centers decreases, it becomes more challenging to provide efficient buffer management to meet the requirements of heterogeneous traffic. We observe that typical shared buffer management policies have two steps: first, they identify short flows arriving at ports and then allocate more buffer room for these ports. Unfortunately, the lack of isolation between long and short flows leads to increased queue buildup and even packet loss of short flows. To address this limitation, we propose D2T, which uses different queue length thresholds for long and short flows. Specifically, we first design a compact data structure to distinguish between long and short flows. Then when two kinds of flows coexist at the same port, the threshold of long flows will decrease to absorb the bursty short flows. What’s more, we introduce D2T${}^{*}$which combines D2T with advanced DRL techniques to move toward mastering buffer management for further improving performance across various scenarios. We implement D2T at a P4-programmable switch and large-scale simulations. The results demonstrate that D2T reduces both average and tail flow completion times (FCT) of short flows by up to 29% and 62% compared with the state-of-the-art policies, respectively.

Large-scale sewage treatment plants are one of the typical Industrial Internet of Things systems, where the presence of a large number of sensors generates massive dynamic time series data, and such multivariate time series data are usually time-dependent and random. Therefore, there is a certain risk when fitting the potential anomalies of real-world data, which will bring great challenges to anomaly detection. In this article, we propose a time-series mutual adversarial network (TMAN), a novel reconstruction model for anomaly detection on multivariate time series. It is based on the idea of adversarial learning and consists of two identical subnetworks. During the training process, two subnetworks can independently complete the learning of the time distribution of normal samples of industrial time series data for mutual adversarial. In the process of detecting, we obtain the residual values of TMAN reconstructed for different time series samples to discriminate anomalies. We combine TMAN and anomaly determination mechanisms to build a new industrial time series anomaly detection framework named TMANomaly. In addition, we select the dataset features with a grey correlation algorithm to achieve very high performance with a small number of features. Experimental results show that our proposed TMANomaly outperforms five popular anomaly detection methods and effectively improves the accuracy of industrial multivariate time series anomaly detection.

In lossless data center networks (DCN), many existing congestion control schemes are used to address the impact caused by priority-based flow control (PFC), such as congestion spreading, and victim flow problems. However, in some special cases, this problem is not solved. Through observation, we examine the interaction between flow control and congestion control, and realize that the mismatch between hop-by-hop flow control and end-to-end congestion feedback, as well as inaccurate rate regulation, are the root causes of frequent PFC triggering. Therefore, we propose Egress Queue Congestion Information Notification (EQCIN). EQCIN implements threshold-based flow identification to avoid packet buildup due to congestion spreading being considered as the root cause of congestion, while using direct feedback from the congestion side to reduce unnecessary link loss. For different flow identifiers, EQCIN adopts different algorithms to achieve targeted rate control. Experimental results show that EQCIN can reduce the number of PFC PAUSEs tends to zero, compared to TIMELY, DCQCN, DCQCN+TCD and improve the link utilization by 7%-77%, respectively.

In mobile edge computing (MEC), task offloading and resource allocation are two important issues that are inextricably linked. However, existing studies have either ignored the mobility of mobile users (MUs) during task offloading or the allocation of profits between two parties during the allocation of limited resources (i.e., the resource competition). In this paper, we jointly optimized these two problems. First, to reduce the task offloading delay and the service interruption due to movement, we develop a mobility-aware model, based on which we propose the MWBS algorithm to select the appropriate offloading base station (BS) for MUs. Second, considering the resource competition and the delay constraint of the task, we develop a double auction model and then propose the DARA algorithm, which efficiently allocates the BS resources and maximizes the total system revenue (i.e., social welfare) through a multi-session auction. Finally, we combine MWBS and DARA to propose the BS resource allocation algorithm called MD-BSRA in mobile scenarios. Simulation results show that MD-BSRA can effectively improve task offload success rate, total system revenue and resource utilization while reducing offload delay and service interruption.

Traditional flowlet routing is a traffic distribution-based network routing algorithm that avoids packet disorder problems. In dynamic network environments, however, this approach may lead to suboptimal performance and network congestion. In this paper, we design a multi-agent flowlet level routing optimization (MAFRO) framework that combines multi-agent deep reinforcement learning (MADRL) and graph neural networks (GNN). MADRL uses multiple agents that interact with the network environment and learn through a trial-and-error process to optimize flowlet routing. Meanwhile, MADRL uses the properties of GNN to model the network topology, interacting and capturing the complex relationships between network nodes. MAFRO enables agents to make more informed and adaptive routing decisions based on the current state of the network, leading to better end-to-end delay and packet loss rate performance. Experimental results demonstrate that MAFRO achieves better performance than the baseline algorithm.

Multi-path TCP (MPTCP) is increasingly popular with the widespread usage of multihomed devices. MPTCP allows data streams to be delivered across multiple simultaneous connections, providing higher bandwidth aggregation and throughput in comparison with single-path TCP. However, due to the path heterogeneity and packet losses, the occurrence of Out-of-Order (OFO) packets is inevitable for MPTCP. Although many approaches have been proposed to mitigate OFO, most of them focused on compensating path delay differences but not considered the impact of packet loss. In this paper, we take the first step towards analyzing the impact of packet loss on OFO, and propose Loss-Aware Throughput Estimation scheduler, LATE. LATE comprehensively considers each subflow's path characteristics and protocol parameters including Round Trip Time (RTT), congestion window (cwnd), and loss rate, to predict the data amount that can be sent over each subflow at a given time and determine wisely which segments should be allocated to which subflows. Experimental results show that LATE achieves a gain of 5.13% in mean goodput with long-lasting flows while reducing the completion time of short flows by about 26.68% compared to the state-of-the-art scheduler for MPTCP.

Multipath transmission control protocol (MPTCP) allows a TCP connection to operate across multiple paths simultaneously and becomes highly attractive to support the emerging mobile devices with various radio interfaces and to improve resource utilization as well as connection robustness. The existing multipath congestion control algorithms, however, are mainly loss-based and prefer the paths with lower drop rates, leading to severe performance degradation in wireless communication systems, where random packet losses occur frequently. To address this challenge and improve the performance of MPTCP in wireless networks, this paper proposes a new mVeno algorithm, which makes full use of the congestion information of all the subflows belonging to a TCP connection in order to adaptively adjust the transmission rate of each subflow. Specifically, mVeno modifies the additive increase phase of Veno so as to effectively couple all subflows by dynamically varying the congestion window increment based on the receiving ACKs. The weighted parameter of each subflow for tuning the congestion window is determined by distinguishing packet losses caused by random error of wireless links or by network congestion. We implement mVeno in a Linux server and conduct extensive experiments both in test bed and in real WAN to validate its effectiveness. The performance results demonstrate that compared with the existing schemes, mVeno increases the throughput significantly, achieves load balancing, and can keep the fairness with regular TCP.