This paper discusses how optical technologies can bring innovation to intra data center networks. The most promising technique is the electrical-packet/optical-circuit hybrid switching network. Its key advantage stems from using large port-count optical switches, which enables flat and unified interconnects among ToRs in a data center; an architecture not possible with electrical switches due to the limited switch chip radix. Such optical switches can be implemented in a way similar to the widely deployed ROADM/OXC that leverages wavelength routing and space switches. When matched to the radix of merchant silicon switches for ToR, the parallel configuration of optical switches yields virtually blocking-free networks, and satisfactory link utilization, say more than 0.6, is expected. With regard to the key issue of optical switch control, the proposed configuration enables decentralized and asynchronous control and the latency is shown to be less than 20 $\mu$ s (L=200m), which is a deterministic value without the long tail as seen in electrical packet switch networks. The presented hybrid switching network will pave the way to future data center networks by resolving the problems expected with the envisaged demise of Moore’s law.

In this paper, we propose a machine-learning-based novel dynamic performance control method for routers that supports several performance levels. The method utilizes support vector machine (SVM) to determine performance level change points. We utilize a traffic normalization technique with a corresponding performance threshold that allows us to apply the same SVM to different traffic volumes. This technique shortens the learning process for control. Several experiments on real Internet traffic data sequences prove that the method yields higher energy efficiency than conventional methods, that is, our previously proposed frequency decomposition-based method and a conventional traffic prediction method based on auto-regressive moving average model with optimal parameter values given by Akaike's information criterion. We also evaluate the impact of several key parameters such as traffic measurement interval and the number of packet processing engines. The results clarify the proper ranges of parameter values that attain significant power reductions while keeping packet loss to acceptable levels.

A novel coarse granular routing scheme for elastic optical networks is proposed in this paper together with a node architecture and network design algorithm. The proposed scheme bundles optical paths to be routed together and each bundle of paths, named flexible waveband, is routed as an entity. Path bundling is done by a small port count wavelength-selective switch (WSS) while flexible waveband routing is done by the other optical switches, i.e., two stage routing. The proposed network design algorithm resolves the routing and frequency slot assignment problem while considering specific constraints imposed by the routing scheme. Numerical experiments on several topologies confirm that the routing performance degradation caused by the coarse granular routing is small while the number of WSSs is substantially reduced.

The six papers in this special issue are devoted to the topic of next generation spectrum efficient and elastic optical transport systems.

We propose a quasi-dynamic network design algorithm considering traffic growth in multi-layered photonic networks that consist of electrical paths and optical paths. The quasi-dynamic network design proposed herein leaves existing paths unchanged in order to minimize service disruption due to path rerouting as well as human error. Simulations verify that the network cost strongly depends on the network design algorithms adopted and the holding terms of service traffic. We, then, introduce a new parameter that represents expected service holding terms. The parameter is shown to represent effectively the network cost trend versus the number of the incremental designs. Furthermore, lower network cost is obtained by taking advantage of the network design algorithms that consider service holding times. This study is the first to address the criteria in designing future multi-layered photonic networks that offer cost-effective expansion while accommodating different services.

We propose a hierarchical optical path network design algorithm. In order to efficiently accommodate wavelength paths in each waveband path, we define a source-destination Cartesian product space that allows the 'closeness' among wavelength paths to be assessed. By grouping 'close' wavelength paths, found by searching for clusters in the space, we iteratively create waveband paths that efficiently accommodate the wavelength paths. Numerical experiments demonstrate that the proposed algorithm offers lower total network cost than the conventional algorithms. The results also show that the hierarchical optical path network is effective even when traffic demand is relatively small.

In this paper, we propose a Wavelet-based prediction method of the Internet traffic volume. By introducing maximal overlap formulation of the Haar wavelet, the proposed method is free from so-called boundary condition, which arises from processing delay of analysis filters of the wavelet transform and refrains from full utilization of information on recent input signals. The proposed method is based on a vector autoregressive model so as to introduce inter-scale correlation of Wavelet coefficient series at different scales.

Dealing with the explosive increase in the amount of Internet traffic requires high-speed and huge capacity Internet protocol (IP) backbone networks. Existing IP backbone networks are constructed using point-to-point wavelength-division-multiplexing (WDM) transmission systems, where all the wavelengths are terminated link-by-link, so that rather expensive optical/electrical conversions are necessary at every node. In these systems, since every IP packet is routed at each intermediate node based on the header information, a header processing bottleneck will occur when the node input traffic exceeds several hundreds of gigabits per second. In order to mitigate these problems, an optical cross-connect (OXC) function that employs wavelength routing of the optical paths (OPs) will provide an effective solution. This paper proposes a network design method where electrical and photonic multiprotocol label switching (MPLS) technologies are used; the network is referred to as a photonic IP network. We first propose new algorithms that minimize the network cost in a multilayered network comprising electrical label switched paths (LSPs) and optical LSPs (optical paths that are controlled using the MPLS mechanism). The particular point of the proposed algorithms is that they include different cost minimization scenarios appropriate for the different OLSP provisioning conditions that are chosen as the first step in the design stage. The effectiveness of the proposed algorithms and the benefits of the OLSPs are quantitatively evaluated through various simulations.

Multiple fiber ring (MFR) networks employing the optical path (OP) concept are discussed to realize robust networks. The architecture is superior to that discussed in several testbeds, with respect to fault recoverability. Both the OP based and non-OP based MFRs can develop several architectures, each of which provides different performance regarding network resource utilization. The required node system scale can be a good measure in determining the network performance. Network design simulations, varying path distribution patterns, have been performed and they elucidated the criteria for selecting the most suitable architecture in terms of resource utilization.

To realize a large scale and robust photonic transport network, this paper proposes a network protection strategy and maintenance signaling flow for WDM optical path transport networks. The network protection strategy is applicable for mesh-type WDM optical path transport networks, and the OA&M flow that is required for the network to recover from failures are clarified. The proposed OA&M flow does not rely on the OP-AIS (optical path alarm indication signal) generation method. This signaling method can also be applied to the all-optical type network.

This paper demonstrates our previously developed large-capacity optical path cross-connect (OPXC) system. The major hurdles to be cleared are (1) establishment of a highly modular architecture that facilitates hardware design and upgrading and (2) the development of an OPXC demonstrator designed to achieve 320-Gbit/s throughput capacity. Due to the use of planar lightwave circuit technologies, compact size packaging for an 8/spl times/16 delivery-and-coupling type optical switch and an arrayed-waveguide grating for a wavelength-demultiplexer are achieved. The dense packaging of the four-channel optical regenerators (3-R function regenerator) makes it possible to realize a large-capacity OPXC that can offer high quality transmission and ensure robustness in terms of multiple node connection. Performance test results confirm the validity of the system design and the feasibility of 320-Gbit/s OPXC implementation.

This paper proposes a new layered transport network architecture on which the WDM optical path network can be effectively created. The optical path network will play a key role in the development of the transport network that will realize the bandwidth-abundant B-ISDN. This paper extends the layered transport network architecture described in ITU-T Recommendation G.803 which is applied in existing SDH networks. First, we elucidate an application example of WDM optical path networks. Next, we propose a new layered architecture for WDM-based transport networks that retains maximum commonality with the layered architectures developed for existing B-ISDN networks. The proposed architecture is composed of circuit layer networks, electrical path layer networks, optical layer networks, and physical media (fiber) networks. The optical layer is divided into an optical path layer and an optical section layer. The optical path layer accommodates electrical paths. Optical section layer networks are divided into optical multiplex section (OMS) layer networks and optical repeater section (ORS) layer networks. The OMS layer network is concerned with the end-to-end transfer of information between locations transferring or terminating optical paths, whereas the ORS layer is concerned with the transfer of information between individual optical repeaters. Finally, a detailed functional block model of WDM optical path networks, the function allocation of each layer, and an optical transport module (OTM) are developed.

This paper proposes optical path accommodation design algorithms for both the wavelength path (WP) and virtual wavelength path (VWP) schemes with or without failure restoration. The optical path (OP) employs WDM/FDM (wavelength/frequency division multiplexing) and wavelength routing in the path layer of the transport network. The proposed algorithms adopt some heuristic approaches to the algorithmic complexity caused by the NP-completeness of the wavelength assignment problem. The objective function is taken as minimizing the average total number optical path cross-connect (OPXC) ports, in other words, OPXC system scale. Utilizing the developed algorithms, extensive simulations quantitatively evaluate the OPXC system scale requirements of a large scale network whose physical topology and path demand distribution pattern are arbitrary. The simulations reveal the basic difference between the WP and VWP schemes.

This paper proposes a layered network architecture for optical transport networks and the concept of optical path network management. First, the WDM (wavelength division multiplex) based optical path network concept and realization techniques such as the WP (wavelength path) and the VWP (virtual WP) are presented. Second, a new layered architecture for WDM based transport networks is proposed. The proposed network architecture retains maximum commonality with existing B-ISDN networks. Finally, optical transport network management functions such as configuration, performance and fault management, are discussed.

The optical path (OP) technology, which employs both wavelength-division multiplexing and wavelength routing, will be the key to enhanced network integrity and an ubiquitous broadband integrated services digital network (B-ISDN) in the future. To construct the OP network, path accommodation design that can solve simultaneously the problems of path routing and wavelength assignment must be established. Since optical wavelengths are scarce resources, even with state-of-the-art technologies, the available number of wavelengths that can be multiplexed into a fiber is restricted to a relatively small number. This entails space division multiplexing with multiple fibers in the link in order to accommodate the large number of paths within a link. This paper proposes novel OP accommodation design algorithms that can heuristically establish wavelength paths (WPs) or virtual wavelength paths (VWPs) in the network, where each link is composed of multiple fibers. These algorithms minimize the average number of fibers (in other words, cross-connect ports) handled at the WP/VWP cross-connect nodes and enable us to obtain the required OP cross-connect (OPXC) system scale at each node in WP/VWP networks. Algorithms that consider failure restoration are also proposed. Some WP/VWP accommodation designs over a polygrid network are simulated using the proposed algorithms. The difference between the WP and VWP schemes in terms of the required OPXC system scale with and without considering failure restoration is quantitatively evaluated for the first time.

The performance limitation of the "leaky bucket algorithm" is analyzed for usage parameter control and traffic management in asynchronous transfer mode (ATM) networks. Simulation results show that the conventional statistical bandwidth allocation method, which uses "the worst pattern derived from the cell interarrival time moments" permitted by the leaky bucket algorithm, can not guarantee the quality of service (QOS). As a result, this paper proves that the VC/VP bandwidth allocation method based on the leaky bucket algorithm is unsatisfactory.>

This paper proposes self-healing network techniques suitable for ATM networks in order to realize a high-reliablity B-ISDN. First, the characteristics of the virtual paths (VP) and their influence on failure restoration are discussed. A high-speed restoration technique which exploits the benefits of the VP is then proposed and described. The technique simplifies the message transmission processes and reduces the number of generated messages by using preassigned backup virtual paths. Next, the scheme used to design the backup VP routes and spare resource distribution for each link is proposed in order to create a network that applies the proposed restoration scheme. Next, self-reconstruction techniques of backup virtual paths are proposed for the realization of a reversionless restoration cycle. Finally, the feasibility of the distributed control operation is discussed.>

Path layer technologies will play a key role in the development of a powerful and failure resilient B-ISDN. So far, they have been based on electrical technologies. This paper highlights WDM/FDM techniques and demonstrates that optical paths can greatly enhance the path layer capability and, therefore, the network performance. It is also shown that effective network failure restoration can be achieved with optical paths. The applicability of the wavelength path (WP) technique to global area networks is revealed by comparing different optical path realization techniques. WPs are applied to the national backbone network example to evaluate the required number of wavelengths, and to identify optical cross-connect node requirements. It then proposes a new optical path concept: the virtual wavelength path (VWP). In the VWP scheme, wavelengths are assigned on a link-by-link basis. In other words, the wavelength assigned to a wavelength path has only local significance. Significant benefits of the VWP such as the simplified path accommodation design within a transmission facility network and the reduced number of wavelengths needed, are elucidated. An optical cross-connect node architecture that enables the VWPs is also proposed. The architecture allows the VWP concept to be realized with commercially available optical technologies. The optical path layer concept proposed exploits and consolidates the layered transport network architecture and optical technologies, and will open up new opportunities for creating a B-ISDN that is bandwidth abundant and has a high degree of integrity.>

Provides an overview of the special issue of the IEEE Journal of Communications, Volume 12, Number 1 (January 1994) which addresses open questions in network integrity, reliability and survivability. Current progress in this area is discussed. The questions addressed include user survivability perspectives on standards, planning, and deployment; the analysis and quantification of network disasters; survivable and fault-tolerant network architectures and associated economic analyses; and techniques to handle network restoration as a result of physical damage or failures in software and control systems. Special interests are devoted to the survivability of broadband networks employing the new transport/switching techniques based on the synchronous optical network (SONET) and asynchronous transfer mode (ATM) standards due to their emerging role in future B-ISDN. Network integrity due to failures of common channel signaling (CCS) systems is also very critical. The present special issue includes 22 papers and is organized into the following sections: user perspectives and planning, software quality and reliability, network survivability characterization and standards, network restoration for SONET networks, network restoration for ATM networks, traffic effect and performance enhancement for computer networks, and survivable network design methods. Network restoration methods for SONET, ATM, and computer networks correspond to those for the physical layer (SONET), ATM layer, and the network layer, defined in the CCITT broadband ISDN layer structure.>

The cell multiplexing characteristics of an asynchronous transfer mode (ATM) network should be evaluated on the so-called worst-case traffic as characterized by the usage parameter control algorithm. A bandwidth allocation method suitable for any periodic cell traffic is proposed. Numerical assessment shows that the proposed method can achieve relatively high link utilization. Using the proposed method, the performances of traffic management schemes adopting the sliding window or leaky bucket algorithms are compared. The result is that traffic control using the sliding window algorithm ensures efficient network resource utilization. Traffic control using the leaky bucket fails to identify the worst-case traffic permitted by the leaky bucket algorithm.

The requirements of broadband asynchronous transfer mode (ATM) audio/video (A/V) transmission systems and the approach used to implement the Washington University broadband terminal (BBT) are discussed. The goal of the project was the development of a system with real-time audio and video capability that fully utilized the capability of the Washington University ATM network. The BBT transmits standard NTSC color video and audio over the local ATM network, using a novel approach. A composite analog A/V signal is digitized at three times the video color burst frequency at the transmitter; at the receiver, a unique frame buffer scheme is used to regenerate the transmitted audio and video signals.>

A credible introduction strategy for B-ISDN is described. The first step is the introduction of asynchronous transfer mode (ATM) techniques into the path layer. The strategy is shown to be effective, since it enables not only the realization of a cost-effective and flexible network infrastructure, but also the provision of high-speed multimedia leased line services. The service advantages possible with virtual paths are also demonstrated. The authors' technical advances, including ATM network resource management techniques and network utilization enhancement techniques, are highlighted.>

The virtual path concept has several valuable features to construct an economical and efficient asynchronous transfer mode (ATM) network. One of them is bandwidth control which affords transmission efficiency improvement through statistical sharing of capacity. An effective bandwidth control algorithm and its calculated performance are described. Network performance with the algorithm is evaluated, and the bandwidth control is shown to successfully improve network transmission efficiency with only a slight increase in processing load compared to the fixed bandwidth scheme. A method is also proposed to equalize call loss probability for each virtual path. The effectiveness of the method is demonstrated by analysis.>

A path and link capacity design method for asynchronous transfer mode (ATM) networks using statistical multiplexing of cells is proposed. By developing the design method, the statistical cell multiplexing effect can be exploited to significantly reduce required network resources. The necessary design procedures are identified. A simple and effective policing mechanism is proposed. An analytical method for the evaluation of cell multiplexing characteristics using declared parameters is provided. Path and link capacity design algorithms that use these techniques are proposed. The effectiveness of the proposed design procedure is confirmed by comparing analytical results to those from a computer simulation.>

A new speech coding and multiplexing scheme matched to the asynchronous transfer mode is described. A block coding technique that is based on a variable-rate coding algorithm that makes the most of the burstiness of voice information is employed. The main feature of the scheme is considerable bit reduction, which is attained by a fairly simple algorithm. It is demonstrated that the proposed algorithm exhibits better quality than that of a 32 kb/s ADPCM at a mean bit rate of less than 13 kb/s. The effect of statistical multiplexing is verified by means of simulation employing long conversational speech samples. Methods for constructing variable- and fixed-length frames (units of information multiplexed and transferred in the network) are proposed. The proposed coding algorithm is shown to be applicable to both variable- and fixed-length frame strategies.>

Broadband transport techniques and network architectures based on the virtual path concept are examined. ATM (asynchronous transfer mode) techniques, when coupled with recent technological innovations, are expected to pave the way for future universal transport networks. The virtual path concept, which exploits the ATM's capabilities, is proposed to construct an efficient and economic network. The concept matches current and anticipated technological trends well. Characteristics and implementation techniques of virtual paths are discussed. Advantages of the virtual path concept and its impact on the transport network architecture are demonstrated. The virtual path strategy is also shown to provide efficiently for networks with dynamic reconfiguration capability which will enhance network performance. Some basic analytical results on the dynamic control effects of virtual paths are provided.>

An exploration is made of some techniques which will play important roles in the development of the future universal transport network. In the network, the information can be efficiently integrated and transported by making the most of the burstiness of the information flows and of the store-and-forward process at the transport nodes. First, a coding and multiplexing scheme for audio signals matched to the above transportation mode is proposed. The effectiveness of the scheme is demonstrated through computer simulation and the extent of the deviations are provided for some sets of parameters. The analysis provides useful tools to perform sensitivity analysis of network delay performance.>

Investigations are carried out on the performance of video transmission systems employing square wave frequency modulation (SWFM). A simple formula for the signal-to-noise ratio (SNR) in the SWFM video transmission systems is derived. The SNR performance is compared to that of frequency modulation (FM) and pulse frequency modulation (PFM). SWFM is shown to display the greatest receiver sensitivity when the fiber 6 dB bandwidth is less than 100 MHz. Video transmission experiments demonstrate the advantages of the SWFM, including 1) degree of simplicity in modulation/demodulation circuits as well as in optical transmitter and receiver circuits, 2) tolerance to nonlinearities in the system, and 3) insensitivity to the fiber baseband frequency-phase characteristics. Both the theoretical and experimental results have made clear the superiority of SWFM optical video transmission systems in application to CATV distribution systems and subscriber loop systems.

Investigations are carried out on the SNR performance in fiber optic FM transmission systems which use multimode fibers and 1.3 μm laser diodes. Both laser diode noise and modal noise are quantitatively evaluated in systems with and without employing a superimposed pulse modulation scheme. Based on these investigations, parameters for analog-digital hybrid signal transmission systems employing FM are determined. The hybrid signal consists of audio, video, and 2 Mbit/s digital data. The results obtained in transmission experiments satisfied the design objectives and confirmed the technical feasibility of an analog-digital hybrid transmission employing FM. The obtained results make the fiber optic FM transmission system with a 1.3\mumLD an attractive alternative in subscriber systems, and video transmission systems for microwave IF links and CATV trunks.

This paper investigates the effects of speckle noise on analog video transmission systems using semiconductor laser diodes. The system linearity degradation due to speckle noise is examined using different fiber types. Then this paper proposes a new modulation technique employing a superimposed pulse scheme to reduce speckle noise. It is experimentally confirmed that the proposed modulation method is effective in reducing speckle noise. Based on the above investigations, analog video transmission experiments are performed using single-mode fibers, step-index multimode fibers, and graded-index multimode fibers in the 0.8 and 1.3 μm wavelength regions. The results of the transmission tests have confirmed the feasibility of analog video transmission using semiconductor laser diodes.