Scalable Traffic Engineering for Higher Throughput in Heavily-loaded Software Defined Networks

Traffic engineering in hybrid SDN networks with multiple traffic matrices

Wide spreading of computer networks along with increasing traffic demand throughout the Internet caused a dramatic increase in energy consumption by networking devices and Internet infrastructure. Energy-aware traffic engineering is a promising approach towards green networking to achieve a trade-off between energy saving and network utilisation in backbone networks. In this paper, we propose to use non-dominated sorting genetic algorithm (NSGA-II) for energy-aware intra-domain traffic engineering. This algorithm tries to make a trade-off between maximum link utilisation (MLU) and energy reservation. For each pair of network topology and traffic matrix, NSGA-II computes the optimal set of links to put to sleep so that the resulting topology would be able to carry the traffic demand. We developed a simulator to evaluate the performance of our mechanism. The results of comprehensive evaluations show that our energy-aware TE approach increases the network performance in terms of energy conservation by 50% at the cost of slight increase in maximum link utilisation.

Traffic engineering (TE) has long been used by network providers to reduce network congestion and improve resource utilization. Due to its significance, several traffic engineering algorithms have been proposed in literature. However, most of these algorithms optimize maximum link utilization (MLU) in network, and/or assume that network has the capability to route demands on arbitrary paths. Optimizing only for MLU can result in longer route computations to save bandwidth along shorter paths, thereby hurting application performance (as shown by recent research). Further, minimizing MLU can lead to solutions where several links have utilization close to MLU, while many others are under- utilized. Besides, as large fraction of today's Internet uses OSPF routing protocol, it cannot benefit from TE algorithms assuming arbitrary routing capabilities. To address these problems, we present Wise-OSPF (WOSPF), a traffic engineering solution for OSPF networks. WOSPF formulates TE as an optimization problem. The objective of WOSPF is to minimize the difference between the maximum and minimum link utilizations across the network, which leads to more uniform traffic distribution compared to optimizing MLU. As WOSPF uses OSPF for routing demands, it does not compute unnecessarily long routes and can be employed in legacy OSPF networks with minimal changes. Our results show that WOSPF reduces standard deviation of link utilizations in network by 31.35% compared to an optimal MLU based TE approach, while achieving an MLU within 1.9% of the optimal.

Traffic Engineering (TE) is an important network operation that routes/reroutes flows based on network topology and traffic demands to optimize network performance. Recently, new emerging applications pose challenges to TE with dynamic network conditions, where frequent routing updates are required to maintain good network performance with Software-Defined Networking (SDN). However, flow rerouting operations could lead to considerable Quality of Service (QoS) degradation and service disruption, which is often neglected by existing TE solutions. In this paper, we apply a new QoS metric named network disturbance to measure the negative impact of flow rerouting operations performed by TE. To achieve near-optimal load balancing performance and mitigate network disturbance together in dynamic network scenarios, we propose a flexible and disturbance-aware TE solution called FlexDATE that combines Reinforcement Learning (RL) and Linear Programming (LP). Specifically, FlexDATE leverages RL to intelligently identify flexible numbers of critical flows for each traffic matrix and reroutes these critical flows based on LP optimization to improve network performance with low disturbance. Empowered by a customized actor-critic architecture coupled with Graph Neural Networks (GNNs), FlexDATE can generalize well to unseen traffic scenarios and remain resilient to single link failures. Extensive simulations are conducted on five real-world network topologies to evaluate FlexDATE with real and synthetic traffic traces. The results show that FlexDATE can achieve the performance target (i.e., 90% of optimal performance) in 99% of network scenarios and effectively mitigate the average and maximum network disturbance by up to 9.1% and 38.6%, respectively, compared to state-of-the-art TE solutions.

This paper proposes a Traffic Engineering (TE) scheme for routing in a target network topology that supports OpenFlow using a metric called congestion contribution and Software-Defined Networking (SDN). According to the network traffic volume has been increasing dramatically, it is clear that the congestion, which is caused by the imbalance of the traffic flow, can be occurred. Furthermore, an Interior Gateway Protocol (IGP) that is used for an Autonomous System (AS) also can cause congestion because it uses snapshot information of network topology to make the shortest path-based routing table. Due to the reasons mentioned above, much research on TE has been conducting to enhance the network performance. However, as an example of TE, the Multi-Protocol Label Switching (MPLS) using Resource Reservation Protocol (RSVP) is not flexible in reconfiguring the labeled path, and it has resource overhead that cannot be ignored. Thus, we propose a TE scheme using SDN to overcome the limitations in the target network. In order to show the feasibility of the proposed scheme, TE application predicts the bandwidth utilization for network devices in SDN networks. It also determines the near-optimized routing path for the maximization of the average bandwidth utilization with avoiding the congestion nodes. Our scheme can configure network devices (e.g., SDN switches) along the paths of traffic flows so that the SDN network can have maximum bandwidth utilization. Through experiments, it is shown that our proposed scheme can utilize the average bandwidth of the whole topology up to 66%.

Traffic matrix (TM) is a critical information for network operation and management, especially for traffic engineering (TE). Due to the technical and mercantile problems, real time measurement for TM is difficult in large scale networks. In this paper, we focus on predicting TM for dynamic traffic engineering. We propose several TM prediction methods based on Neural Networks (NN) and predict TM from three perspectives: predict the overall TM directly, predict each origin-destination (OD) flow separately and predict the overall TM combined with key element correction. In addition to the prediction accuracy, we evaluate different prediction methods through the performance of TE, as well as the prediction time. We test the proposed methods by real world datasets from Abilene, CERNET and GÉANT. The experiment results show that prediction methods based on Recurrent Neural Networks (RNN) can achieve better prediction accuracy than methods leveraging Convolutional Neural Networks (CNN) and Deep Belief Networks (DBN). Predicting each OD flow through RNN models can further improve the prediction accuracy, as well as the performance of TE under the OSPF network scenario, the SDN/OSPF hybrid network scenario and the multi-commodity flow problem scenario. However, it takes longer prediction time for predicting each OD flow sequence. In contrast, predicting the overall TM combined with key element correction can provide a trade-off between the TE results and the prediction overhead, which is more appropriate for dynamic TE in the above three scenarios.

Software Defined Networking (SDN) is an emerging network architecture that separates the control plane from the data plane to simplify and improve network management with a high degree of flexibility. Network optimization under traffic uncertainties is one of the most challenging topics in communication networks to optimize network performance and traffic delivery. Although the traffic optimization technology has been extensively studied in the industry, a traffic optimization solution different from the traditional network is needed in the SDN network, which can utilize global network information and traffic characteristics to control and manage traffic in a better way. In this paper, a Mixed Linear Geometric Programming Traffic Optimization Algorithm (MLGP-TOA) is proposed for the problem of traffic uncertainties in SDN. Aiming at minimizing the maximum link utilization (MLU), the initial problem is transformed into a convex optimization problem by monomial approximation and variable substitution. Then, the inner point method is used to find the global optimal solution, and the optimal split ratio at each node is obtained. Finally, the configuration information is sent to the data plane. The simulation results show that the algorithm can reduce MLU, so that the traffic can fully utilize network resources and avoid congestion.

Software-defined networking (SDN) apps are developed to support various functions (e.g., traffic engineering, routing, security, etc.) for SDN networks. Their operations rely on the APIs offered by the control plane. They may be compromised or designed to be malicious by third parties. Though there have been many studies against malicious apps, they only restrict the APIs used by them with coarse-grained controls. In this work, we seek to show that some malicious flow entries cannot be detected or prevented by current defenses. They may impede the operations of control-plane services or hinder packets from being forwarded correctly in the data plane. To show their negative impact, we devise two attacks, topology spoofing and forwarding-based DoS, as well as examine their damage and analyze root causes. We then propose a context-aware, event-based anomaly detection (CEAD) framework to defend against the malicious flow entries. It provides more fine-grained controls over the flow entries set by apps. Different from other studies, it does anomaly detection by examining the context correlation between an event, the app registering it, and the flow entries set by the app for the event. Our evaluation results show that the CEAD can detect all the malicious flow entries in our given cases, and confirm its scalability with negligible overhead at increasing TCP connection attempt rates.

Open Shortest Path First (OSPF) is one of the most widely used intra-domain routing protocol. It is well known that OSPF protocol does not provide flexibility in terms of packet forwarding to achieve any network optimization objective. Because of the high cost of network assets and commercial and competitive nature of Internet service provisioning, service providers are interested in performance optimization of their networks. This helps in reducing congestion hotspots and improving resource utilization across the network, which, in turn, results in an increased revenue collection. One way of achieving this is through Traffic Engineering. Currently traffic engineering is mostly done by using MPLS. But legacy networks running OSPF would need to be upgraded to MPLS. To achieve better resource utilization without upgrading OSPF network to MPLS is a challenge. In this paper we present a simple but effective algorithm, called Smart OSPF (S-OSPF) to provide traffic engineering solution in an OSPF based best effort network. We formulate an optimization problem based on the traffic demand to minimize the maximum link utilization in the network. Routing of the traffic demand is achieved using OSPF. We have simulated S-OSPF on real networks of two service providers. Simulation results show that S- OSPF based traffic engineering solution performance very closely follows the optimal solution.

Many traffic engineering and network design tasks require a network-wide description as an input for their performance evaluation. Therefore, a traffic matrix represents the demands for these tasks and is applicable to them. Traffic matrix can be measured directly or estimated from an indirect data. Generating synthetic traffic matrix method, is a better way to achieve traffic matrix indirectly. We have previously proposed the conqueror traffic based synthetic traffic matrix and we have indicated that our conqueror traffic based matrix performed better in all cases where conqueror traffic existed. In this paper by measuring the maximum link-utilization, we consider how well the conqueror traffic based synthetic matrix optimizes the specific traffic engineering task; that is optimizing network routing to minimize congestion. First, we optimize the routing with the conqueror traffic based synthetic matrix to minimize congestion. Then, we test the performance of the resulting routing on the real traffic matrix. Simulation results indicate that OSPF optimization which we use in this paper to optimize the weights based on the conqueror traffic based synthetic matrix, optimize the routing better and result in a better performance on real traffic matrix.

Software‐defined networking (SDN) is revolutionizing network architecture by offering incredible scalability, flexibility, and programmability. SDN's applications have expanded into various domains, including cloud computing, edge computing, enterprise networks, data centers, and the Internet of Things. One of its standout features is its ability to enhance traffic engineering. Traffic engineering involves optimizing and dynamically controlling traffic flows to reduce congestion and minimize the maximum link utilization, all while managing network traffic from a centralized point. SDN boosts overall network performance, improves user experience, and ensures that resources are used effectively. Minimizing maximum link utilization means distributing network traffic across different links or paths to make use of the most available resources and avoid congestion. This process is often referred to as load balancing or congestion minimization within the network. The primary aim is to prevent any single link from becoming overloaded. Over the years, various algorithms have been proposed to tackle this issue, but still such algorithms struggle with effectively minimizing the maximum link utilization. The proposed work achieves approximately 24% and 16% improvement in load balancing; it reduces congestion by 94% and 60% when compared to tabu search and enhanced tabu search.

Software Defined Network (SDN) has been envisioned as the next generation network infrastructure, which simplify network management by decoupling the control plane and data plane. It is becoming the leading technology behind many traffic engineering solutions, since it allows a central controller to globally plan the paths of the flows. However, Ternary Content Addressable Memory (TCAM), as a critical hardware storing rules in SDN-enabled devices, can be supplied to each device with very limited quantity because it is expensive and energy-consuming. To efficiently use TCAM resources, we address the routing rule space occupation problem for multiple unicastSessions with Quality-of-Service (QoS) constraints. To our best knowledge, this is the first work to joint routing rule optimization with traffic engineering for multipath flows. We formulate the problem using Mixed Integer Linear Programing (MILP) and propose an approach based on ant colony algorithm to solve it. Finally, we evaluate the expected performance of our proposed algorithm through a simulation study.

Inter-AS outbound Traffic Engineering (TE) aims to control the flow of traffic exiting an AS so as to optimize inter-AS TE objectives such as load balancing among multiple downstream ASes. The inter-AS traffic matrix is fundamental input for outbound TE and its accuracy is essential for the target performance to be achieved. However, deriving an accurate traffic matrix is far from trivial. This paper proposes a robust approach for outbound TE to manage traffic demand uncertainty through Scenario-based Robust Optimization. The objective of this robust outbound TE is to minimize the worst-case maximum inter-AS link utilization across a set of inter-AS traffic matrices while minimizing the performance deviation from the optimal solutions. Simulation results reveal that the robust outbound TE is capable of achieving reasonably good performance under all the given traffic matrices while non-robust approaches are not.

Ternary Content Addressable Memory (TCAM) is widely used for caching flow entries in Software Defined Network (SDN). However, limited by the current technology level and hardware cost, the capacity of TCAM is difficult to meet the needs for caching flow entries of large scale SDN, which would seriously affect the scalability of SDN. Aimed at the problem, some researches improve the efficiency of TCAM capacity by setting different lifecycle for different flow entries. Such as Openflow protocol proposed the default flow entry lifecycle, AHTM scheme dynamicly adjusted flow entry lifecycle according to the number of interrupted flows. However, these schemes didn't consider the dynamic features of network traffic in SDN, and hardly conform to the actual network demands. In order to effectively utilize TCAM capacity and thus enhance the scalability of SDN, we propose TF-IdleTimeout scheme to dynamically set flow entry lifecycle according to real-time network traffic in SDN-based data center network. Two criterion, called Flow Entry Missing Number and Flow Dropping Number, are used to evaluate the effect. The results show that, compared with existing schemes, TF-IdleTimeout scheme can greatly improve the utilization efficiency of TCAM capacity in SDN.

In recent years, software-defined networking (SDN), which performs centralized network management with software, has attracted much attention. In SDN, the data plane is separated from the control plane. Network administrators can dynamically change the setting of the data plane and control routing on the control plane through a control protocol such as OpenFlow. Although packets are transmitted based on flow entries in the data plane, the number of flow entries that data plane can have is limited. It is difficult to perform flexible routing control with a small number of flow entries. To overcome this difficulty, this paper proposes a routing method that reduces the number of flow entries, while reducing the maximum link utilization. The proposed method allocates common paths to flows that can be aggregated, considering the utilization of the links. Through numerical experiments, this paper shows the effectiveness of the proposed method.