Estimation of end-to-end network traffic plays an important role in traffic engineering and network planning. The direct measurement of a network's traffic matrix consumes large amounts of network resources and is thus impractical in most cases. How to accurately construct traffic matrix remains a great challenge. This paper studies end-to-end network traffic reconstruction in large-scale networks. Applying compressive sensing theory, we propose a novel reconstruction method for end-to-end traffic flows. First, the direct measurement of partial Origin-Destination (OD) flows is determined by random measurement matrix, providing partial measurements. Then, we use the K-SVD approach to obtain a sparse matrix. Combined with compressive sensing, this partially known OD flow matrix can be used to recover the entire end-to-end network traffic matrix. Simulation results show that the proposed method can reconstruct end-to-end network traffic with a high degree of accuracy. Moreover, in comparison with previous methods, our approach exhibits a significant performance improvement.

A Compressive Sensing-Based Approach to End-to-End Network Traffic Reconstruction

Network virtualization is a relatively new research topic A number of articles propose that certain benefits can be realized by virtualizing links between network elements as well as adding virtualization on intermediate network elements In this article we argue that network virtualization may bring nothing new in terms of technical capabilities and theoretical performance, but it provides a way of organizing networks such that it is possible to overcome some of the practical issues in today?s Internet We strengthen our case by an analogy between the concept of network virtualization as it is currently presented in research, and machine virtualization as proven useful in deployments in recent years First we make an analogy between the functionality of an operating system and that of a network, and identify similar concepts and elements Then we emphasize the practical benefits realized by machine virtualization, and we exploit the analogy to derive potential benefits brought by network virtualization We map the established applications for machine virtualization to network virtualization, thus identifying possible use cases for network virtualization We also use this analogy to structure the design space for network virtualization

Network virtualization: a hypervisor for the Internet?

The introduction of new services requiring large and dynamic bitrate connectivity can cause changes in the direction of the traffic in metro and even core network segments throughout the day. This leads to large overprovisioning in statically managed virtual network topologies (VNTs), which are designed to cope with the traffic forecast. To reduce expenses while ensuring the required grade of service, in this paper we propose a VNT reconfiguration approach based on data analytics for traffic prediction (VENTURE). It regularly reconfigures the VNT based on the predicted traffic, thus adapting the topology to both the current and the predicted traffic volume and direction. A machine learning algorithm based on an artificial neural network is used to provide robust and adaptive traffic models. The reconfiguration problem that takes as its input the traffic prediction is modeled mathematically, and a heuristic is proposed to solve it in practical times. To support VENTURE, we propose an architecture that allows collecting and storing data from monitoring at the routers and that is used to train predictive models for every origin-destination pair. Exhaustive simulation results of the algorithm, together with the experimental assessment of the proposed architecture, are finally presented.

/pdf/virtual-network-topology-adaptability-based-on-data-5acmyrj5me.pdf

Virtual Network Topology Adaptability Based on Data Analytics for Traffic Prediction

One approach to accommodating traffic on a wavelength-routed optical network is to construct a virtual network topology (VNT) by establishing a set of lightpaths between nodes. To accommodate fluctuating traffic on a VNT, we propose an adaptive VNT control method, which reconfigures VNTs according to traffic conditions on VNTs, in IP over wavelength-routed wavelength-division-multiplexing networks. To achieve adaptability in the VNT control method, we focus on attractor selection, which models behaviors where biological systems adapt to unknown changes in their surrounding environments and recover their conditions. The biological system driven by attractor selection adapts to environmental changes by selecting attractors at which the system condition is preferable. Our VNT control method uses deterministic and stochastic behaviors and controls these two appropriately by simple feedback of the conditions of an IP network. By utilizing stochastic behavior, our new approach adapts to various changes in traffic demand with selecting suitable attractors, which correspond to VNTs in our method, for the current traffic demand. Moreover, to define feedback of the conditions on the IP network, our proposed scheme only uses load information on links, which is easily and directly retrieved and thus achieves quick responses to changes in traffic demand. The simulation results indicate that our VNT control method based on attractor selection quickly and adaptively responds to various changes in traffic demand, and our method adapts to at most twice larger changes in traffic demand than existing heuristic approaches.

Adaptive Virtual Network Topology Control Based on Attractor Selection

Network virtualization is an emerging trend claimed to reduce the costs of future networks. The key strategy in network virtualization is of slicing physical resources (links, routers, servers, etc.) to create virtual networks composed of subsets of these slices. One important challenge on network virtualization is the resource management of the physical or substrate networks. Sophisticated management techniques should be used to accomplish such management. The sophisticated techniques offered by autonomic communications rise as an appropriated alternative to address the challenges of managing the efficient use of substrate resources on network virtualization. Thus, this paper proposes a distributed self-organizing model to manage the substrate network resources. An evaluation scenario is depicted and simulations show that approximately 36.8% of the network traffic load can be spared when the self-organizing model is enabled in the evaluated scenario.

Distributed autonomic resource management for network virtualization

Traffic matrix is essential to traffic engineering (TE) methods. Because it is difficult to monitor traffic matrices directly, several methods for estimating them from link loads have been proposed. However, estimated traffic matrix includes estimation errors which degrade the performance of TE significantly. In this paper, we propose a method that reduces estimation errors while reconfiguring the virtual network topology (VNT) by cooperating with the VNT reconfiguration. In our method, the VNT reconfiguration is divided into multiple stages instead of reconfiguring the suitable VNT at once. By dividing the VNT reconfiguration into multiple stages, our traffic matrix estimation method calibrates and reduces the estimation errors in each stage by using information monitored in prior stages. We also investigate the effectiveness of our proposal using simulations. The results show that our method can improve the accuracy of the traffic matrix estimation and achieve an adequate VNT as is the case with the reconfiguration using the actual traffic matrices.

Gradually reconfiguring virtual network topologies based on estimated traffic matrices

In this paper, we present a practical VNT (virtual network topology) reconfiguration method for large-scale IP and optical networks with traffic matrix estimation considerations. We newly introduce a partial VNT reconfiguration algorithm with multiple transition stages. By dividing the whole VNT transition sequence into multiple transitions, estimation errors are calibrated at each stage by using network state information of prior stages. Because estimation errors are mainly due to the fewer information in the estimated traffic matrix calculation, our approach tries to increase the constraint conditions for traffic matrix estimation by introducing partial reconfiguration, and to relax the impact of estimation errors by limiting the number of optical-paths reconfigured at each stage. We also investigate the effectiveness of our proposal through simulations and clarify the robustness against estimation errors by using partial reconfiguration.

Gradually Reconfiguring Virtual Network Topologies Based on Estimated Traffic Matrices

The growth of the Internet and emerging application layer technologies causes numerous changes in network environments. Therefore, it becomes important to achieve adaptive methods of controlling networks in addition to optimizing their performance. To achieve an adaptive network control method, we focus on attractor selection, which models behaviors where biological systems adapt to unknown changes in their surrounding environments and recover their conditions. In this paper, we show the applicability of the attractor selection to the adaptive virtual network topology (VNT) control in IP over wavelength-routed WDM networks. The simulation results indicate that our VNT control method based on attractor selection quickly and adaptively responds to various changes in traffic demand.

Application of Attractor Selection to Adaptive Virtual Network Topology Control

Application of attractor selection to adaptive virtual network topology control

One approach for accommodating IP traffic on the WDM (Wavelength Division Multiplexing) network is to construct a logical topology by establishing a set of lightpaths between nodes. The lightpath carries IP traffic and does not require any electronic processing at intermediate nodes, which reduce the load of packet processing at those nodes. However, if IP and WDM networks have its own routing functions, the lightpaths configured in WDM network may not be fully utilized by IP unless the route of IP packets can be explicitly determined. Therefore, the integration of routing mechanisms is necessary to provide efficient resource utilization and flexible adaptation against traffic changes. In this paper, we propose an integrated routing mechanism for IP over WDM networks. The key idea is to prepare a set of virtual-links based on a resource utilization of the WDM network, and then calculate the minimum cost route on the IP network. Our simulation results show that the performance of our method outperforms those of optimally designed logical topologies when traffic patterns change. Fluid flow network simulation is conducted for simulating IP traffic flow on WDM networks.

Shin'ichi Arakawa

Papers

Gradually reconfiguring virtual network topologies based on estimated traffic matrices

Gradually Reconfiguring Virtual Network Topologies Based on Estimated Traffic Matrices

Application of Attractor Selection to Adaptive Virtual Network Topology Control

Application of attractor selection to adaptive virtual network topology control

On the integration of IP routing and wavelength routing in IP over WDM networks