International Journal of Network Management最新文献

The rapid growth of vehicular applications has resulted in high demand for Internet technology, which demands an unprecedented network capacity and a high quality of service (QoS). In vehicular ad hoc networks (VANETs), since nodes (vehicles) are highly mobile. The dynamic nature of the network topology in the VANET system changes due to frequent changes in link connectivity. The vehicles-to-vehicles (V2V), vehicles-to-infrastructure (V2I), and QoS, as well as the heterogeneity of applications within the VANET. VANETs have been introduced to make driving comfortable by providing safety and support to drivers. Due to the flexibility and offloading schemes available in-vehicle applications, there are some limitations. However, there are many issues in providing optimum service provisioning and scheduling in the vehicular environment. In VANETs, BSs and roadside units (RSUs) improve QoS. However, Internet services transmit packets to vehicles using stochastic models, and it predicts the traffic on a VANET. We provide open challenges to drive stochastic models in this direction.

The classification of network traffic, which involves classifying and identifying the type of network traffic, is the most fundamental step to network service improvement and modern network management. Classic machine learning and deep learning methods have widely adopted in the field of network traffic classification. However, there are two major challenges in practice. One is the user privacy concern in cross-domain traffic data sharing for the purpose of training a global classification model, and the other is the difficulty to obtain large amount of labeled data for training. In this paper, we propose a novel approach using federated semi-supervised learning for network traffic classification, in which the federated server and clients from different domains work together to train a global classification model. Among them, unlabeled data are used on the client side, and labeled data are used on the server side. The experimental results derived from a public dataset show that the accuracy of the proposed approach can reach 97.81%, and the accuracy gap between the federated learning approach and the centralized training method is minimal.

In software-defined networking (SDN), the controller relies on the information collected from the data plane for route planning, load balancing, and other functions. Statistics information is the most important kind of information among them, so the correctness of statistics information is the key to the proper operation of the network. Most of the current research on data plane focuses on policy consistency, rule redundancy, forwarding anomalies, and so on, and little attention is paid to whether the statistics information uploaded by the switches to the controller is correct. However, incorrect statistics information inevitably leads the controller to make wrong decisions. Therefore, this paper proposes an audit-based malicious information correction mechanism to address the problem of wrong statistics information uploaded by the switches. This mechanism audits the statistics information and locates malicious switches before uploading the statistics information to the controller. It identifies and corrects the statistics information errors by combining flow path and statistics information. We have performed simulations on Nsfnet, Abilene, and Fat-Tree, and the results show that our method can correct about 70% of the statistical information errors with less computational cost. To the best of our knowledge, this paper is the first malicious statistics information correction scheme for wildcard rules.

Recently, the virtual data center embedding (VDCE) problem has drawn significant attention because of a growing need for efficient means of data center resource allocation. By ensuring a set of virtual data center (VDC) integration requests coming from his customers, among the main concern of an infrastructure provider is the maximization of the utilization rate of data center resources and benefits. However, existing VDCE solutions mostly focus on consolidating virtual machines in a single physical data center. Therefore, in this work, we improve the consolidated targets techniques, that consider only the virtual machines integration, by the consideration of network devices and fabrics (e.g., switches and paths/links). We consider new unreleased constraints such as multiple virtual nodes of the same request co-location, and intermediate node requirements when a virtual link is mapped. To address the above problem, in this paper, we propose a binary linear programming-based model, called BLP-VDCE, to solve the VDCE problem with network-aware consideration. This model ensures a simultaneous consolidated embedding of virtual nodes and virtual links. Extensive simulations show that solving the proposed BLP-VDCE model can efficiently embed VDC requests with a high physical resource utilization rate.

Recent years have witnessed the rise of novel network applications such as telesurgery, telepresence, and holoportation. As such applications have stringent performance requirements, timely and accurate traffic monitoring becomes of paramount importance to be able to react in a timely and efficient manner, and swiftly adjust the network configuration to achieve the sought-after requirements. However, existing monitoring schemes are either incurring high cost (e.g., high bandwidth consumption) due to the large number of monitoring messages or inefficient when they incur high reporting delay (i.e., the time needed for a monitoring message to reach the controller) making the collected statistics obsolete. In this paper, we address this problem and propose monitoring mechanisms for software defined networks that minimize the monitoring cost while satisfying an upper bound on the reporting delay of the statistics. Our solutions allow to carefully select the switch that should report the statistics about each flow crossing the network taking into consideration the available bandwidth and the capacity of the switch (i.e., the maximum number of flows that it can monitor). In particular, we formulate the switch-to-flow selection problem as an integer linear program and propose two heuristic algorithms to cope with large-scale instances of the problem. We consider the scenario where a single controller is collecting statistics and another where statistics are collected by multiple controllers. Simulation results show that the proposed algorithms provide near-optimal solutions with minimal computation time and outperform existing monitoring strategies in terms of monitoring cost and reporting delay.