Distance-adaptive modulation is effective at enhancing network capacity as it allows the maximum possible modulation order to be selected for each optical path. However, present single-carrier systems can only select just one modulation order for each optical path and hence the adaptability to transmission characteristics is strictly limited. In contrast, digital subcarrier multiplexing systems can select a combination of modulation orders for multiple subcarriers on each optical path and can flexibly adapt to various transmission characteristics. This paper numerically evaluates the transmission characteristics of digital subcarrier multiplexing systems. The interaction between laser phase noise and chromatic dispersion is well examined by extensive simulations, and two-phase estimation methods are compared. The results show that digital subcarrier multiplexing systems with the appropriate phase estimation method enable longer transmission distances.
The popularity of high-capacity communication services such as video streaming and cloud computing has accelerated the growth in IP traffic. In order to effectively manage and maintain networking systems, various intelligent technologies based on software-defined networking (SDN) have been widely studied. An SDN system that offers flexible optical path management exploiting optical performance monitoring, digital signal processing, and resource allocation is expected to realize higher capacity networks by lowering margins needed to offset system uncertainty. In this paper, we provide a comprehensive survey of optical path management schemes based on machine learning.
To satisfy the ever-increasing traffic demands in a cost-effective manner, the transmission capacity per wavelength must be increased. The use of high-order modulation and high-baud modulation can increase the capacity per wavelength. However, such signal forms are susceptible to IQ impairments created in the transceiver since symbol distance and symbol duration are extremely short. The IQ impairments can be alleviated by adaptive finite-impulse-response (FIR) filters in the receiver. However, transmitter IQ phase imbalance and transmitter IQ power imbalance cannot be eliminated completely in the receiver-side digital signal processing (DSP). Furthermore, to eliminate transmitter IQ skew, additional filters need to be implemented in the receiver DSP. Therefore, transmitter IQ impairments should be calibrated at the transmitter side to enhance demodulation performance and cost efficiency. In this paper, we propose a novel method for measuring transmitter IQ impairments. In our proposed scheme, transmitter IQ impairments are calculated from the tap coefficients of the adaptive filters adopted in the measurement system. In addition, our proposed scheme can estimate transmitter IQ impairments even when other impairments contaminate the signal. With the estimated value, transmitter-side DSP can easily eliminate the transmitter IQ impairments with digital buffers. Simulations using 64 Gbaud 64-QAM signals show that the estimation error of IQ phase imbalance is less than 1.5 degree, that of IQ power imbalance is less than 0.2 dB, and that of IQ skew is less than 0.05 ps even if the signal is contaminated by other impairments.
Digital coherent reception is an attractive candidate for increasing the channel capacity without expanding the bandwidth needed. The receiver DSP circuit must be able to compensate for the IQ-phase/power mismatch with low computation power requirements. In this paper, we propose a novel DSP circuit suitable for short-reach transmission systems including intra-datacenter networks. The proposed DSP circuit offers polarization-mode demultiplexing, compensation for IQ-phase/power mismatch, and estimation of carrier phase and frequency offset. The proposed DSP circuit consists of three-stage one-tap FIR filters; therefore, complicated calculations are not required. Its good demodulation performance is confirmed by simulations.
The introduction of quasi-Nyquist wavelength-division multiplexing (WDM) reduces unused frequency resources and hence attains high spectral efficiency. Applying quasi-Nyquist WDM to optical-path networks using wavelengthselective switches (WSSs) is hindered by two factors: the limited bandwidth resolution of WSS passbands and the significant spectrum narrowing occasioned by WSS traversal. In this paper, we propose a novel quasi-Nyquist WDM network architecture, where the bundle-based wavelength assignment and receiver-side partial-response spectrum shaping are simultaneously utilized so as to resolve the two problems. Extensive computer simulations show that the proposed network architecture increases the maximum attainable node-hop count compared to the conventional transmission systems.
We propose a cost-effective metro network architecture with fiber-granular routing and path-granular add/drop operations together with its ILP-based design algorithm. The proposal alleviates the impact of filtering impairment while using already deployed OXC/ROADM nodes. Numerical simulations on several real-world metro topologies verify that it increases the spectral efficiency compared to the ideal method for DWDM networks.
Optical-path networks based on wavelength-selective switches (WSSs) can cost-effectively process wavelength-divisionmultiplexed (WDM) signals. To deal with the continuously increasing network traffic, the spectral efficiency must be improved by minimizing guardband bandwidths. Quasi-Nyquist WDM systems are seen as offering the highest spectral efficiency. However, such highly dense WDM systems suffer from signal-spectrum narrowing induced by the nonrectangular passbands of WSSs. Furthermore, widely deployed WSSs cannot process quasi-Nyquist WDM signals since the signal-alignment granularity does not match the passband resolution of the WSSs. In this paper, we propose a network architecture that enables quasi-Nyquist WDM networking. First, multiple channels are bundled so that the total channel bandwidth matches the WSS-passband resolution. Second, the number of spectrum-narrowing events of each path is limited by our restriction-aware algorithm. These proposals allow a 100-GHz bandwidth to accommodate three 100-Gbps DP-QPSK signals aligned with 33.3-GHz spacing and a 200-GHz bandwidth to accommodate three 400-Gbps dual-carrier DP-16QAM signals aligned with 66.6-GHz spacing. Intensive network analyses confirm that the spectral efficiency is improved by up to 46.4%. Feasibility is verified by transmission experiments using 69-channel 400-Gbps dual-carrier DP-16QAM signals aligned with 66.6-GHz spacing in the extended C-band. The fiber capacity of 27.6 Tbps and the transmission distance of 800 km are attained by our proposed quasi-Nyquist WDM networking.
The traffic volume processed within the datacenter increases exponentially. In a typical datacenter, top-of-rack (ToR) switches connected to servers are interconnected via multi-stage electrical switches. The electrical switches require power-consuming optical-to-electrical and electrical-to-optical conversion. To resolve the problem, a single optical switch needs to be introduced to offload large-capacity flows. The optical switch in the datacenter must have a large number of input/output ports to support many ToR switches. The combination of delivery-and-coupling (DC) switches and wavelength-routing (WR) switches comprised of 1xN non-cyclic arrayed-waveguide gratings (AWGs) can attain high-port-count switches. To further increase the port count, the system loss must be reduced or higher-power transmitters must be used. To overcome this difficulty, we propose novel optical-switch architecture in which nxN uniform-loss and cyclic-frequency (ULCF) AWGs are utilized for the WR-switch part, where the system loss can be reduced by the factor of n. To confirm the effectiveness of our proposal, 12x48 ULCF AWGs were newly fabricated with planar-lightwave-circuit (PLC) technology. Part of a 1,536x1,536 optical switch was constructed, and good transmission performance was experimentally confirmed by bit-error-ratio measurements in 96-wavelength 32-Gbaud DP-QPSK signals in the full C-band. The throughput was 153.6 Tbps.
We analyze the maximum transmission distance and hop count of M-QAM signals, where link and node transmission characteristics are jointly considered. With the modulation format optimally determined by the analyses, spectral efficiencies of ultra-dense wavelength-division-multiplexing (WDM) networks are maximized.
Ever-increasing intra-datacenter traffic will spur the introduction of high-baud rates and high-order modulation formats. Increasing symbol rates and modulation levels decreases tolerance against transmission impairment that includes chromatic dispersion. Transmission distance in warehouse-scale datacenters can be several kilometers, and then management of chromatic dispersion is necessary. Dispersion-compensating fibers are widely deployed in backbone networks, however, applying them in datacenters is not cost-effective since wavelength channels are coarsely multiplexed. In digital coherent systems, signal distortion due to chromatic dispersion can be resolved in digital domain; however, it will take long time before coherent systems can be introduced in datacenter networks because of their high cost. In this paper, we propose a novel impairment mitigation method employing machine learning. The proposed method is effective even after non-coherent detection and hence it can be applied to cost-sensitive intra-datacenter networks. The machine learns optimum symbol-decision criteria from a sequence of dispersed training signals, and it discriminates payload signals in accordance with the established decision criteria. With the scheme, the received signals can be demodulated in the presence of large chromatic dispersion. The transmission distance thus can be extended without relying on costly optical dispersion compensation. Since information of transmission links is not a priori required, the proposed scheme can easily be applied to any datacenter network. We conduct transmission experiments using 400-Gbps channels each of which comprises 8-subcarrier 28-Gbaud 4-ary pulse-amplitude-modulation (PAM-4) signals, and confirm the effectiveness of the proposed scheme.
We propose introducing aligned frequency assignment to each bandwidth channel, i.e., the semi-flexible grid, with the goal of minimizing frequency slot fragmentation under dynamic flexible-grid network expansion considering the expected need for channel capacity upgrades. In semi-flexible grid networks, for each set of channels having the same frequency bandwidth, a regular grid is defined where its spacing is the same as the required bandwidth and channels are aligned to their corresponding grids. A network expansion algorithm is developed that maximizes the efficiency of semiflexible grid assignment to achieve efficient channel bandwidth upgrading. Numerical experiments prove that the number of fibers necessary and the degree of fragmentation in the frequency domain are reduced by 15% and 80%, respectively, compared to conventional flexible grid networks accommodating the same traffic.
Intra-datacenter traffic is growing more than 20% a year. In typical datacenters, many racks/pods including servers are interconnected via multi-tier electrical switches. The electrical switches necessitate power-consuming optical-to- electrical (OE) and electrical-to-optical (EO) conversion, the power consumption of which increases with traffic. To overcome this problem, optical switches that eliminate costly OE and EO conversion and enable low power consumption switching are being investigated. There are two major requirements for the optical switch. First, it must have a high port count to construct reduced tier intra-datacenter networks. Second, switching speed must be short enough that most of the traffic load can be offloaded from electrical switches. Among various optical switches, we focus on those based on arrayed-waveguide gratings (AWGs), since the AWG is a passive device with minimal power consumption. We previously proposed a high-port-count optical switch architecture that utilizes tunable lasers, route-and-combine switches, and wavelength-routing switches comprised of couplers, erbium-doped fiber amplifiers (EDFAs), and AWGs. We employed conventional external cavity lasers whose wavelength-tuning speed was slower than 100 ms. In this paper, we demonstrate a large-scale optical switch that offers fast wavelength routing. We construct a 720×720 optical switch using recently developed lasers whose wavelength-tuning period is below 460 μs. We evaluate the switching time via bit-error-ratio measurements and achieve 470-μs switching time (includes 10-μs guard time to handle EDFA surge). To best of our knowledge, this is the first demonstration of such a large-scale optical switch with practical switching
time.
We propose a novel optical network architecture that uses waveband virtual links, each of which can carry several optical paths, to directly bridge distant node pairs. Future photonic networks should not only transparently cover extended areas but also expand fiber capacity. However, the traversal of many ROADM nodes impairs the optical signal due to spectrum narrowing. To suppress the degradation, the bandwidth of guard bands needs to be increased, which degrades fiber frequency utilization. Waveband granular switching allows us to apply broader pass-band filtering at ROADMs and to insert sufficient guard bands between wavebands with minimum frequency utilization offset. The scheme resolves the severe spectrum narrowing effect. Moreover, the guard band between optical channels in a waveband can be minimized, which increases the number of paths that can be accommodated per fiber. In the network, wavelength path granular routing is done without utilizing waveband virtual links, and it still suffers from spectrum narrowing. A novel network design algorithm that can bound the spectrum narrowing effect by limiting the number of hops (traversed nodes that need wavelength path level routing) is proposed in this paper. This algorithm dynamically changes the waveband virtual link configuration according to the traffic distribution variation, where optical paths that need many node hops are effectively carried by virtual links. Numerical experiments demonstrate that the number of necessary fibers is reduced by 23% compared with conventional optical path networks.
An effective solution to the continuous Internet traffic expansion is to offload traffic to lower layers such as the L2 or L1 optical layers. One possible approach is to introduce dynamic optical path operations such as adaptive establishment/tear down according to traffic variation. Path operations cannot be done instantaneously; hence, traffic prediction is essential. Conventional prediction techniques need optimal parameter values to be determined in advance by averaging long-term variations from the past. However, this does not allow adaptation to the ever-changing short-term variations expected to be common in future networks. In this paper, we propose a real-time optical path control method based on a machinelearning technique involving support vector machines (SVMs). A SVM learns the most recent traffic characteristics, and so enables better adaptation to temporal traffic variations than conventional techniques. The difficulty lies in determining how to minimize the time gap between optical path operation and buffer management at the originating points of those paths. The gap makes the required learning data set enormous and the learning process costly. To resolve the problem, we propose the adoption of multiple SVMs running in parallel, trained with non-overlapping subsets of the original data set. The maximum value of the outputs of these SVMs will be the estimated number of necessary paths. Numerical experiments prove that our proposed method outperforms a conventional prediction method, the autoregressive moving average method with optimal parameter values determined by Akaike’s information criterion, and reduces the packet-loss ratio by up to 98%.
We propose a novel optical path routing mechanism that combines coarse-granularity optical multicast with fine-granularity add/drop and block. We implement the proposal in an optical cross-connect node with broadcast-and-select functionality that offers high cost-effectiveness since no addition equipment from conventional ROADMs is needed. The proposed method, called branching, enhances the routing capabilities over the original grouped routing networks by enabling wavelength paths to be established through different GRE pipes. We also present a novel path/GRE routing and wavelength/GRE index assignment algorithm that supports the new routing function. Numerical experiments using real network topologies verify the improved routing performance and the superior efficiency of the proposed control algorithm over original GRE-based networks.
With the continuous increase in Internet traffic, reconfigurable optical add-drop multiplexers (ROADMs) have been widely adopted in the core and metro core networks. Current ROADMs, however, allow only static operation. To realize future dynamic optical-network services, and to minimize any human intervention in network operation, the optical signal add/drop part should have colorless/directionless/contentionless (C/D/C) capabilities. This is possible with matrix switches or a combination of splitter-switches and optical tunable filters. The scale of the matrix switch increases with the square of the number of supported channels, and hence, the matrix-switch-based architecture is not suitable for creating future large-scale ROADMs. In contrast, the numbers of splitter ports, switches, and tunable filters increase linearly with the number of supported channels, and hence the tunable-filter-based architecture will support all future traffic. So far, we have succeeded in fabricating a compact tunable filter that consists of multi-stage cyclic arrayed-waveguide gratings (AWGs) and switches by using planar-lightwave-circuit (PLC) technologies. However, this multistage configuration suffers from large insertion loss and filter narrowing. Moreover, power-consuming temperature control is necessary since it is difficult to make cyclic AWGs athermal. We propose here novel tunable-filter architecture that sandwiches a single-stage non-cyclic athermal AWG having flatter-topped passbands between small-scale switches. With this configuration, the optical tunable filter attains low insertion loss, large passband bandwidths, low power consumption, compactness, and high cost-effectiveness. A prototype is monolithically fabricated with PLC technologies and its excellent performance is experimentally confirmed utilizing 80-channel 30-GBaud dual-polarization quadrature phase-shift-keying (QPSK) signals.
This paper describes an inter-domain path control system for large-scale photonic networks, especially considering
applicability to multiple carrier domain networks. The overall architecture of multi-domain optical network control
system is reviewed. Our system employs PCE (Path Computation Element) based routing model and utilizes maximum
flow information in path computation and Path Key scheme in signaling, which improve the load-balancing and the
confidentiality of the inside information for the route computation and the path provisioning among different carriers.
The measured performance results for inter-domain path computation and path setups using global optimization scheme
is also provided. It includes estimation of the path provisioning time at the network scale dependency with the route
computation time and the signaling time. As a result, its applicability to a 1000-node scale photonic network is shown.
We propose a novel optical multicast tree design algorithm for hierarchical optical path networks that exploit waveband
routing. To efficiently groom wavelength multicast trees into waveband trees, we define a generalized inclusive relation
between node groups that identifies neighboring destination groups. The algorithm then establishes a minimum weight
waveband tree that contains the wavelength multicast trees that include the neighboring destinations. Numerical
experiments confirm that the proposed waveband multicast can significantly reduce the number of optical switch ports.
We propose a hierarchical optical path network design algorithm that considers wavelength conversion.
Numerical experiments prove that, by employing wavelength conversion, hierarchical optical path networks
will be more cost effective than single-layer optical path networks over a wide range of traffic demand, which is
not possible without using wavelength conversion.
We propose a simple node architecture that is used to concatenate two ROADM rings. The proposed architecture
employs variable filters that can effectively handle classified node input signals to realize the desired routing operations.
Finally, we developed a prototype system and performed transmission experiments. The technical feasibility has been
confirmed.
We evaluate switch scale for several single-layer and hierarchical optical cross-connects. With the introduction of
practical waveband add/drop ratio bound that can be derived through recently developed network design algorithms,
switch scale can be significantly reduced. We identify the most cost-effective architectures by applying practical
parameter values that are determined through our studies.
KEYWORDS: Control systems, Algorithm development, Optical networks, Computing systems, Optimization (mathematics), Time metrology, Visibility, Broadband telecommunications, Fiber to the x, Standards development
This paper proposes an inter-domain path control system based on PCE (Path Computation Element) for large-scale
photonic networks, especially for a photonic network across over multiple carrier domains. We discuss the overall
architecture of multi-domain optical network control system. The system utilizes maximum flow information in path
computation and Path Key scheme in signaling, which improve the load-balancing and the confidentiality of the inside
information for the route computation and the path provisioning among different carriers. We also provide measured
performance results for inter-domain path computation and path setups using global optimization scheme, estimating the
path provisioning time at the network scale dependency with the route computation time and the signaling time. As a
result, we show its applicability to a 1000-node scale photonic network.
We propose a dynamic RWA method for optical path networks that pre-computes route candidates to attain better load-balancing.
Numerical experiments elucidate that the proposed method can accommodate 1.4-1.8 times more traffic than
the conventional method that ignores load-balancing.
We propose a 2-stage ILP-based design algorithm for hierarchical optical path networks that utilize hybrid-HOXCs. The
hybrid-HOXC consists of an optical waveband cross-connect and an electrical cross-connect which grooms only
wavelength paths. Its effectiveness is evaluated through numerical experiments. Impact of electrical/optical port cost
ratio on the total network cost is also investigated.
We propose an inter-domain path routing algorithm for multi-domain photonic networks. The proposed algorithm
introduces a step-wise weighting technique and utilizes averaged link load information of each domain and that of a link
set between each adjacent domain pair. Numerical experiments demonstrate that the proposed algorithm outperforms a
conventional method in terms of blocking probability and matches that achieved without domain segmentation.
We propose an efficient network architecture to implement optical fast circuit switching. Future bandwidth abundant
services such as Ultra High Definition Television (UHDTV), lambda-leased line services, and layer-one optical VPNs
will generate less-bursty traffic that will fill wavelength path capacity. To realize effective optical fast switching
networks, we introduce a hierarchical structure that combines physical network and optical path levels. A higher physical
layer network (transit network) bridges several lower layer networks (local networks). The optical path layer is divided
into two layers; a waveband path (a group of wavelength paths), and wavelength path layer. The transit networks employ
large granular optical paths, waveband paths. The transit network creates an adaptive virtual topology that can efficiently
carry wavelength path connection requests between lower layer network nodes. Numerical experiments show that the
proposed hierarchical network greatly reduces the necessary number of optical switch ports at the blocking probability
equivalent to that of the single layer network. The effectiveness of the proposed architecture are confirmed for various
network sizes.
The hierarchical optical path network that utilizes wavebands is recognized as very important in meeting the future
explosive growth of traffic demand. The use of backup paths is crucial to realize reliable networks. In order to build
survivable hierarchical optical path networks, the two types of protection mechanisms implemented in the optical layer
are identified: waveband protection and wavelength path protection. We have already developed a novel network design
algorithm that utilizes waveband protection and showed that it can reduce network costs significantly. Another type of
protection, wavelength path protection, was also developed and we have demonstrated that further network cost
reduction can be attained in the area of small traffic demands. The effectiveness of the wavelength path protection
algorithm was confirmed for some network parameter values, however, further clarification is necessary regarding the
impact on network cost of network parameters, especially waveband capacity, a major network parameter. This paper
investigates network cost variation with waveband capacity for hierarchical optical path networks that utilize waveband
and wavelength path protection. Numerical experiments demonstrate the importance of waveband capacity optimization.
The recent advances in network control technology such as GMPLS enable us to dynamically manage path connections
with different granularities through a unified control interface, which can improve network resource utilization. In this
paper, we investigate the effect of path granularity and operation interval on dynamic path network operation with the
aim of minimizing unused bandwidth and the frequency of control actions. We develop a simple technique that applies
low-pass filters to the varying Internet traffic; we show that Internet traffic streams captured in different environment
exhibit a common trend with regard to Internet traffic parameters.
We propose a dynamic path allocation algorithm that considers all paths in the network simultaneously over a given time
period. Numerical experiments demonstrate that the algorithm substantially reduces the blocking probability for path
establishment.
We propose a routing and wavelength assignment algorithm for Optical Burst Switching (OBS) networks that utilizes
centralized control. First, a method that can estimate the expected total blocking time in the network is presented. Then
the proposed algorithm minimizes the estimated blocking time by simple iterative local optimization in terms of the
traffic demand between each pair of nodes. We demonstrate that the proposed algorithm attains much smaller blocking
probability than conventional distributed control algorithms. It is also shown that with introduction of optical buffers and
burst retransmission, the proposed method realizes low burst loss rates (<10-6) acceptable for most applications.
This paper proposes new switch architectures for hierarchical optical path cross-connect (HOXC) systems. The architectures allow incremental expansion of system scale in terms of the number of input/output fiber ports, wavebands, and optical paths per waveband. These feature assure the cost-effective introduction of the HOXCs even at the outset when traffic volume is not so large. Furthermore the effectiveness of the proposed switch architectures is demonstrated in a comparison with single-layer OXCs (conventional OXCs). The results provide useful criteria for the introduction of HOXCs.
This paper proposes a segmented network architecture that is very effective in constructing a large scale network that uses transparent OXC's, and clarifies the network design criteria. Network costs with different network design scenarios are investigated. In particular, IP plus optical layer networks with multi-layer traffic engineering and one plus one optical path protection are examined. Numerical experiments show the effects of the segment size on the network cost and that the proposed designs are effective in designing large-scale networks.
KEYWORDS: Optical discs, Objectives, Semiconductor lasers, Signal processing, Monochromatic aberrations, Light sources, Digital video discs, Data storage, Head, Relays
The possibility of an optical disk capacity of 28GB/side is studied. The read-write characteristics of a phase-change optical disk are examined using a thin cover layer, a blue laser diode, high NA objective lens and the PRML method. The wide tilt margins of high-density recording are achieved.
The attainable recording density of a thermally balanced phase-change optical disk with the mark position recording method have been studied, focusing on the viewpoint of designing an optical disk memory system. A new recording method with constant duration of the read power level after every write-pulse, called the off-pulse, was applied. A wide overwrite power margin from -20% to +10% was obtained for a bit length of 0.63 micrometers and a track pitch of 0.9 micrometers . These results mean that it is possible to attain 1.3 GByte capacity in a 90 mm double-sided disk with the zoned constant angular velocity format.
The mark edge recording characteristics of a new thermally balanced phase-change optical disk operating at a wavelength of 690 nm have been studied, especially from the viewpoint of designing an optical disk drive. As a recording method, the multi-pulse write compensation method with constant duration of the read power level immediately after every multi-pulse, called the off-pulse, was applied. A high carrier-to-noise ratio (CNR) of 54 dB, a high erasability of 35 dB, and an overwrite power margin from -15% to +5% were obtained for a bit length of 0.5 micrometers and a track pitch of 1.0 micrometers at 14.7 m/s linear velocity. These results mean that it is possible to design an optical disk drive providing a capacity of 1.1 GBytes using a 90 mm double-sided disk. Further, it has been found that the off-pulse plays a very important role in controlling the precision of mark edge recording.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.