Toru Katagiri

Experimental Demonstration of Disaggregated Emergency Optical System for Quick Disaster Recovery

Disaggregation and white-box approaches are useful in expediting disaster recovery of optical transport networks. When an optical network system is damaged by a large disaster, the damaged devices can be replaced by the disaggregated functional devices, while disregarding vendor boundaries. In the light of the disaggregation and white-box approaches, we are developing a prototype of disaggregated portable emergency optical system (EOS) for early and low-cost postdisaster recovery. The EOS is customizable, and different functions can be selected to meet the different requirements in disaster recovery. In addition to replacing the damaged functions of the original optical system with an EOS, we introduce two new disaggregated functions into the EOS for postdisaster recovery. First, we introduce an optical supervisory channel handshake scheme to aid the interconnection of the surviving optical resources. Second, we introduce a scheme to achieve the quick recovery of the damaged control plane with the surviving or first restored wireless access capability. This is highly desired not only for emergency network control, but also for the quick collection of the network damage information. These two new functions have been implemented into the EOS prototype. We experimentally demonstrated the network recovery of the data-plane and the control-plane with the EOS.

Wavelength defragmentation for seamless service migration

Emerging applications require future elastic optical networks to provide dynamic and reconfigurable services while maintaining efficient usage of wavelength resources. In this paper, we propose integer linear programming (ILP)-based wavelength defragmentation solutions with minimal optical path disruptions during the migration process. The new approaches achieve up to 20% resource efficiency improvement with less than a quarter of disruptions compared to a conventional approach in the sample networks. In addition, we enhance our approach to shorten the computation time and evaluate it under a computation time limitation of 10 min. The enhanced approach also achieves a resource efficiency improvement of more than 20% with no optical path disruption in the realistic networks, compared with the case where defragmentation is not supported.

A new index of hidden workload for firewall rule processing on virtual machine

Network Functions Virtualization (NFV) is a technology for running software-based functions on commodity hardware, which can lead to cost savings and flexible functionality. However, due to the software processing of network functions, the performance of Virtualized Network Functions (VNFs) is considerably degraded in relation to the types of VNFs and the configuration of VNF applications. Therefore, for a practical NFV system, grasping the accurate workload of VNF applications and managing the VNF performance are important. This paper focuses on analysis of a virtual firewall as a representative VNF. We first reveal new insights about the performance characteristics of the virtual firewall through preliminary experiments. Furthermore, we propose a method for estimating the hidden workload of the virtual firewall by using the number of rules in an Access Control List (ACL) and the traffic volume for each rule. Finally, we show the effectiveness of the proposed methods.

Demonstration of an autonomous, software controlled living optical network that eliminates the need for pre-planning

A “living network” is demonstrated, that, unlike todays static planning, adapts to varying network conditions, and allows operation close to performance limits. An autonomously operating test-bed shows feasibility with an allocated margin of only 1.5 dB.

Extension of OpenFlow protocol to support optical transport network, and its implementation

By having the OpenFlow protocol support Optical Transport Networks (OTN) and work as the unified control interface for multilayer (L0-L4) networks, a simple and cost-effective multilayer Software Defined Networking (SDN) controller can be created. Our aim is to propose and standardize the extended OpenFlow protocol to support OTN in Open Networking Foundation (ONF). We describe the approved specification, and the process from proposal to standard. We applied this extension to our OpenFlow controller and OpenFlow agent and succeeded in demonstrating that the extension works as desired in single- and multi-vendor environments. Future directions for OpenFlow extension and modeling of multilayer integrated nodes are also discussed.

Cost-efficient multi-layer network design employing traffic re-aggregation and shared protection across layers

Shared protection/restoration is a promising solution for reducing protection resources and is supported at each layer of the current multi-layer networks. Software-defined networking is expected to reduce equipment cost as well as operational cost by orchestrating these shared protection functionalities. However, although protection resource sharing improves link utilization, it sometimes increases the required equipment. Meanwhile, traffic re-aggregation at each layer is an important technique for low volume traffic to utilize the underlying link capacity more efficiently, but re-aggregation also makes it difficult to share protection resources with traffic at lower layers. In this paper, we present multi-layer network design strategy and method that reduce equipment cost by means of both traffic re-aggregation at each layer and protection resource sharing among multiple service traffic at different layers. The strategy first prioritizes traffic re-aggregation at each layer, and then maximally delegates shared protection to lower layers as long as it does not increase the required capacity at the lower layer. Evaluation results from the example three-layer networks confirm that the proposed method can effectively reduce equipment cost compared to the conventional design method. Cost reduction is achieved by leveraging shared protection functions at multiple layers.

Accurate prediction of quality of transmission based on a dynamically configurable optical impairment model

We have proposed a dynamically configurable and fast optical impairment model for the abstraction of the optical physical layer, enabling new capabilities such as indirect estimation of physical operating parameters in multivendor networks based on pre-FEC BER information and machine learning. BER is commonly reported by deployed coherent transponders; therefore, this scheme does not increase hardware cost. The estimated parameters can subsequently be used to predict optical signal quality at the receiver of not-already-established optical connections more accurately than possible based on the limited amount of information available at the time of offline system design. The higher accuracy and certainty reduce the required amount of required system margin that must be allocated to guarantee reliable optical connectivity. The remaining margin can then be applied toward increased transmission capacity, or a reduced number of regenerators in the network. We demonstrate the quality of transmission prediction experimentally in an optical mesh network with 0.6 dB Q-factor accuracy, and quantify the benefit in terms of network capacity gain in metro networks by impairment-aware network simulation.

Noise-induced VNE method for software-defined infrastructure with uncertain delay behaviors

Abstract Software-defined infrastructure (SDI) provides virtualized infrastructures to customers by slicing computing resources and network resources. One of the important problems for deploying an SDI framework is to control the assignment of physical resources to a virtual network against changes of traffic demand and service demand. For this problem, the virtual network embedding (VNE) problem, which maps a virtual network to physical resources, has been addressed, but a centralized calculation was assumed. It is difficult to adopt the centralized approach as the size of infrastructure increases and the number of VN requests increases because the identification of current demand becomes more complicated. In this paper, we present a VNE method that works with limited information for large, complicated, and uncertain SDI frameworks. To achieve this, our method applies the biological “Yuragi” principle. Yuragi is a Japanese word whose English translation is “a small perturbation to the system.” Yuragi is a mechanism that provides adaptability of organisms and is often expressed as an attractor selection model. This paper develops a Yuragi-based VNE method that deals with node attributes, has the generality to handle a performance objective, and runs in multi-slice environments. Simulation results show that the Yuragi-based method decreases VN migrations by about 29% relative to a heuristic method to adapt to fluctuations in resource requirements.

Towards seamless service migration in network re-optimization for optically interconnected datacenters

Emerging applications require datacenter networks to provide dynamic and reconfigurable services in order to adapt to the dynamicity of cloud traffic, as well as to maintain the optimality of resource utilization. In this paper, we demonstrate novel re-optimization design techniques for realizing agile and seamless service. Our main contribution is a novel integer linear programming (ILP) based approach which can minimize connection disruption, while optimizing resource utilization in a re-optimization process. We propose an ILP-based approach that integrates resource assignment and resource dependency digraph construction. We also propose a method to determine migration order. We compare the proposed ILP-based approach with a heuristic approach in terms of the number of connection disruptions and the migration process. Simulation results demonstrate the effectiveness of our approach.

Wavelength defragmentation with minimum optical path disruptions for seamless service migration

We propose wavelength defragmentation solutions with minimum optical path disruptions. The new approaches achieve up to 20% resource efficiency improvement with less than one fourth of disruptions compared to the conventional approach.