In an Internet Protocol network running a link-state routing protocol, determining the link weights selects each source-to-destination route on which the sum of the link weights is minimized. Previous studies have focused on determining physical link weights to reduce the network congestion ratio in case of physical-link failures. However, no study has yet addressed a model that determines link weights by incorporating traffic-engineering (TE) links and investigates the effect of incorporating TE links on reducing network congestion. The link-state routing protocol treats a TE link as a logical, direct link between nonadjacent nodes. This paper proposes a link-weight design model with TE links based on preventive start-time optimization (PSO) for handling single physical-link failures, called PSO-TE. The model considers physical and TE links when determining the link weights under the assumption of all single physical-link failures. It identifies the set of link weights that minimizes the worst-case network congestion ratio across all considered failure patterns. Introducing TE links does not require additional physical link resources and thus does not increase capital expenditures. Numerical results demonstrate that PSO-TE reduces the worst-case network congestion ratio compared to PSO without TE links. PSO-TE reduces the worst-case network congestion ratio compared with other models, including PSO without TE links, start-time optimization, and inverse capacity weighting.

This paper proposes a priority-aware guaranteed hardware resource allocation in virtual packet optical nodes (VPONs) and describes experimental evidence of service provisioning with the proposed method on testbed. A network based on the VPON brings solution of service diversification and traffic explosion because it provides multiple services by hardware level separated virtual networks such as IP/Ethernet/ Multi-Protocol Label Switching. The VPON has reconfigurable hardware which is dedicatedly allocated for each service and service-specific logical functions are implemented respectively. However, a previous hardware resource allocation method is not efficient. In previous work, the hardware resources once allocated to a service are not reallocated until the service terminated whenever the hardware is not optimally used. In recent years, hardware reconfiguration time has become shorter and it has enabled to apply for service demand change by releasing hardware from services with low demand and reallocating it to other services with high demand. Thus, this paper proposes preemptive dynamic hardware resource reallocation algorithm for VPON and a priority-aware guaranteed hardware resource reallocation method to realize efficient resource usage and high service capacity. Results of the computer simulation show improving service capability. Furthermore, to demonstrate the feasibility of service provisioning with the proposed method, emulators of the virtual packet optical node were constructed and performed resource reallocation. Results of the experiment show that the virtual packet optical node can configure function chains for providing services based on calculation results of the proposed resource reallocation method.

This paper proposes a scheduling model for updating resource allocations for virtualized networks (VNs) without congestion. The proposed model determines the schedule of migrating traffic flows on VNs from old routes to new routes. The model aims to minimize the number of rounds required to complete the update of all existing VNs. We evaluate the performance of model in terms of the percentage of trials where feasible update scheduling exists and the number of rounds required to complete the update. Numerical results show that more rounds are required to achieve congestion-free update when the traffic demand or the number of VNs increases. The number of required rounds tends to remain the same when the maximum number of rounds exceeds a certain value; this observation helps network operators estimate the time required for VN update.

The provisioning cost of a virtualized network (VN) depends on several factors, including the numbers of virtual routers (VRs) and virtual links (VLs), mapping of them on a substrate infrastructure, and routing of data traffic. An existing model, known as the virtual network embedding (VNE) model, determines the embedding of given VN graphs into the substrate infrastructure. When the resource allocation model of the VNE problem is adopted to a single-entity scenario, where a single entity fulfills the roles of both a service provider and an infrastructure provider, an issue of increased costs of VNs and access paths arise. This paper proposes a model for virtualized network graph design and embedding (VNDE) for the single-entity scenario. The VNDE model determines the number of VRs and a VN graph for each request in conjunction with embedding. The VNDE model also determines access paths that connect customer premises and VRs. We formulate the VNDE model as an integer linear programming (ILP) problem. We develop heuristic algorithms for the cases where the ILP problem cannot be solved in practical time. We evaluate the performance of the VNDE model on several networks, including an actual Japanese academic backbone network. Numerical results show that the proposed model designs suitable VN graphs and embeds them according to the volume of traffic demands and access path cost. Compared with the benchmark model, which is based on a classic VNE approach, the proposed model reduces the provisioning cost at most 28.7% in our examined scenarios.

This article proposes a primary and backup resource allocation model that provides a probabilistic protection guarantee for virtual machines against multiple failures of physical machines in a cloud provider to minimize the required total capacity. A physical machine allocates both primary and backup computing resources for virtual machines. When any failure occurs, the survived physical machines with preplanned backup resources recover the virtual machines on the failed physical machines and take over the workloads. The probability that the protection provided by a physical machine does not succeed is guaranteed within a given number. Providing the probabilistic protection can reduce the required backup capacity by allowing backup resource sharing, but it leads to a nonlinear programing problem in a general-capacity case against multiple failures. We apply robust optimization with extensive mathematical operations to formulate the primary and backup resource allocation problem as a mixed integer linear programming problem, where capacity fragmentation is suppressed. We prove the NP-hardness of considered problem. A heuristic is introduced to solve the optimization problem. The results reveal that the proposed model saves about one-third of the total capacity in our examined cases; it outperforms the conventional models in terms of both blocking probability and resource utilization.

Deadline-aware job scheduling problems have been attracting attention in the application domains of scientific workflows and commercial data-center wide area networks. Satisfying all of the demands before the corresponding deadlines requires optimally spreading traffic along future times. In this paper, we propose a novel deadline-aware job scheduling method that leverages a deep reinforcement learning policy gradient algorithm. We obtain the following findings through several experiments on evaluating the proposed method. The EDF-trained initial policy with the proposed method outperforms a random initial policy. We evaluate how the training period influence the performance of the proposed method and the relationship between the training period length and the success ratio. We also demonstrate the proposed method outperformed with other job-scheduling methods, i.e., round-robin, first-come-first-served, and earliest-due-first, with the gaps between these methods and the proposed method expanding as the traffic load increased.

This paper proposes a backup resource allocation model that provides a probabilistic protection for primary physical machines in a cloud provider to minimize the required total capacity. When any random failure occurs, workloads are transferred to preplanned and dedicated backup physical machines for prompt recovery. In the proposed model, a probabilistic protection guarantee is introduced to prevent the cloud provider from capacity overbooking. We apply robust optimization in our model to formulate the backup resource allocation problem as an integer linear programming problem. A simulated annealing heuristic is adopted to solve the same optimization problem when the cloud provider is large. Finally, the results reveal that the required backup capacity depends on the reliability of primary physical machines. Specifically, the more the resources in primary physical machines share backup capacity when the failure probabilities of primary physical machines are sufficiently small, the less capacity is required for backup resource allocation.

In this paper, we propose an optimization model for designing multiple network functions virtualization (NFV)-based campus area networks (CANs). Organizations, such as universities and research institutions, have their own campus information and communication technology (ICT) equipment, and it is desired that this equipment be moved to NFV/cloud data centers of high reliability and resiliency. However, NFV-based CAN is not affordable because the data transmission cost is higher with a public cloud. One solution is for multiple organizations to procure NFV/cloud data center resources together. By doing so, the cost of these resources will be reduced. There are planning issues to solve when choosing optimal NFV/cloud sites in order to make progress on this approach. The proposed model minimizes the total network cost incurred by multiple organizations including the wide area network cost. It is formulated and analyzed by using mixed integer liner programming. The effect of cost minimization was evaluated in a ladder network, and the cost reduced up to 50%. This reduced cost will encourage organizations to deploy NFV-based CANs.

In this paper, we've explored the possibility of expecting/avoiding degradation of network service performance of Network function virtualization (NFV) by long-term monitoring of hardware and network performances, and adjusting resource assignments of VNFs by preventive management based on Neural Networks (NNs). We've constructed an experimental network service, periodically inserted the specific amount of stresses to the hardware and networks systems, and let the NN learn the long-term trends of service performance changes, then try to adjust VNF resource assignment based on NN's expectations. We evaluate and discuss the results of our experiments.

We propose an optimization model for designing multiple network functions virtualization (NFV)-based campus area networks (CANs). Organizations, such as universities and research institutions have their own campus information and communication technology equipment, but many would like to move this equipment to NFV and cloud data centers for improving reliability and resiliency. However, NFV-based CAN is not affordable for them, because costs are higher with a cloud. One solution is for multiple organizations to procure NFV and cloud data center resources together. By doing so, their individual costs of using these resources will be reduced. To make progress on this approach, there are planning issues to resolve when choosing optimal NFV and cloud data center locations. The proposed model minimizes the total network costs incurred by the organizations, including the wide area network cost and data synchronization costs for recovery from faults at data centers and the various subcampus network configurations of legacy CANs. The model is formulated and analyzed by using mixed integer linear programming. The effect of cost minimization is evaluated in a ladder network and an actual network, SINET5, and it is found that the costs can be reduced by up to 63%. The calculation times of this model under practical conditions are short and the model will be useful in practice. It is also shown that the cost of fault recovery can be suppressed. These results will encourage organizations to deploy NFV-based CANs.

We propose a virtualization architecture for multi-campus information and communication technology (ICT) equipment with integrated cloud and NFV capabilities. The aim of this proposal is to migrate most ICT equipment on campus premises into cloud and NFV platforms. Adopting this architecture would make most ICT services secure and reliable and their disaster recovery (DR) economically manageable. We also analyze a cost function and show the cost advantages of this proposed architecture, describe implementation design issues, and report a preliminary experimentation of NFV DR transaction. This architecture would encourage academic institutes to migrate their own ICT systems located on their premises into cloud environments.

SINET5 is a new 100-Gbps-based academic backbone network, which started full-scale operations in April 2016. It uses multi-protocol label switching-transport profile (MPLS-TP) systems and reconfigurable optical add-drop multiplexers (ROADMs) to create a nationwide network and has more than 50 backbone IP routers to provide a wide range of services, such as several virtual private network (VPN) services. It provides end-to-end data communications up to 100 Gbps throughput, minimized-latency, and software-defined networking (SDN)-friendly functions to researchers in every Japanese prefecture. SINET5 is also a platform for dynamic inter-cloud connections and network functions visualization (NFV) services. This paper brief review of the network architecture, and describes new featured services, SDN-oriented layer-2 on-demand VPN services, and NFV functions. Field test results for SINET5 performance are also reported.

We consider a mechanism for providing a multi-reliability service in multilayer GMPLS networks. By introducing GMPLS restoration techniques and shared risk link group management, we can provide a highly reliable protected connection service. However, there is a trade-off between reliability and efficiency of network resource usage. In addition, reliability requirements differ depending on the type of service (e.g., Internet access or leased line). Thus, we propose a mechanism for calculating an efficient route for a protected connection that can satisfy specific reliability conditions requested by customers. We also present simulation results that indicate our schemes are remarkably effective for achieving a better balance between end-to-end reliability and efficiency. We also demonstrate the quantitative relationship between availability of service and resource efficiency through simulation experiments

In this paper, we consider a mechanism for providing path protection in multi-region networks. Here a region includes an interior gateway protocol (IGP) area, an autonomous systems (AS) and a layer network. Many path protection schemes have been proposed, but most of them have been focused on a protection mechanism within a region network. We thus propose a mechanism for providing inter-region protection, which is used for highly reliable carrier services. We also point out inter-region protection schemes are remarkably effective in improving end-to-end reliability. The key to our proposal lies in a multilayer disjoint routing algorithm, called MLD, that enables us to find multi-layer protection paths. We investigated the performance of our scheme through extensive simulations, and the simulation results show that our approach is sure to find a set of failure-independent paths while achieving better utilization of network resources.

We propose a buffer management mechanism, called V-WFQ (virtual weighted fair queueing), for achieving approximately fair bandwidth allocation with a small amount of hardware in high-speed networks. The basic process for allocating bandwidth fairly uses selective packet dropping to compare the measured input rate of the flow with an estimated fair bandwidth share. Though V-WFQ is a hardware-efficient FIFO-based algorithm, it can achieve almost ideal fairness in bandwidth allocation. Simulation results show that V-WFQ achieves a good balance between fairness and link utilization under various simulation conditions.

Next-generation Internet backbone node architectures are proposed that offer sophisticated QoS support including real-time voice, video streaming, and VPN. The backbone core node, named MSN ( multi-service network) Type-X, has 320 Gbit/s throughput and IPv4/v6 dual stack forwarding capability. Note that, through the employment of state-of-the art VLSI technology, both v4 and v6, OC-192C full wire-rate forwarding performance is achieved for the first time. In addition, sophisticated DiffServ with programmable thousands of virtual queues is also proposed to realize a complete set of hardware. In addition, Type-X uses an ATM emulation technique that can into allow the new packet-centric backbone to support conventional services. Finally, the paper describes high-reliability techniques such as newly structured service non-stop software and a firmware upgrade technique. These newly proposed techniques and systems will realize the next-generation, real-time, high-QoS capability, and VPN IP backbone network.

A new scheduling algorithm that concatenates data blocks to increase switching throughput is proposed. The algorithm controls the degree of concatenation and reduces the number of switching instances to improve switch utilization. The switch architecture uses virtual output queue switching architecture where the core switch fabric is an optical matrix switch. The optical matrix switch requires the guard-time overhead needed for optical switch control. By reducing the number of switching instances, guard-time overhead can be reduced and switch utilization can be improved. Computer simulations show that efficiency is dramatically increased and that fairness in terms of data throughput among output ports is achieved.

The Internet was not originally designed to prevent malicious users from securing more resources than other users. Although the best effort service offered by the Internet is quite well known, realizing fair availability in the best effort service is becoming necessary. Accordingly, this paper proposes the maximal network that realizes fair availability. This new concept has two points (1) all users are guided to use some form of (adequate) rate control by the incentives of the proposed maximal network (2) the fair availability between rate controlled users is provided. As guidance, the proposed network penalizes users who refuse to use rate control. Using a distributed algorithm to realize max-min distribution, the proposed network offers fair availability. Its implementation is described and its feasibility is evaluated by computer simulations. It is shown that fair availability can be realized. Lastly, the concept of multi-class best effort service is proposed and simulations show that this service is possible in the maximal network.