Flexible grid optical networks allow a better exploitation of fiber capacity, by enabling a denser frequency allocation. A tighter channel spacing, however, requires narrower filters, which increase linear intersymbol interference (ISI), and may dramatically reduce system reach. Commercial coherent receivers are based on symbol by symbol detectors, which are quite sensitive to ISI. In this context, Nyquist spacing is considered as the ultimate limit to wavelength-division multiplexing (WDM) packing. In this paper, we show that by introducing a limited-complexity trellis processing in the receiver, either the reach of Nyquist WDM flexi-grid networks can be significantly extended, or a higher spectral efficiency (SE) is possible at equal reach. By adopting information-theoretic techniques, we design a limited-complexity trellis processing and quantify its SE gain in flexi-grid architectures where wavelength selective switches over a frequency grid with 12.5GHz granularity are employed.

The Multicanonical Monte Carlo (MMC) technique is a new form of adaptive importance sampling (IS). Thanks to its blind adaptation algorithm, it does not require an in-depth system knowledge for exploitation as does traditional IS. Hence MMC is a practical, handy tool to estimate via simulation the probability of rare events in complex telecom systems, such as the symbol error rate or the outage probability. In this paper, we present the analytical connections between MMC and IS, and describe the recursive algorithm via which MMC seeks an optimal "flat-histogram" warping. We also provide practical guidelines on how MMC can be successfully applied in telecom to achieve accelerations of simulation time by many orders of magnitude with respect to standard Monte Carlo.

This paper provides design guidelines for gain-clamped doped-fiber amplifiers in a wavelength division multiplexed networking environment. A simple dynamic model of the doped fiber amplifier allows us to derive explicit expressions for the small-signal response, which help identify and optimize the most critical parameters, pump power and laser wavelength, for best dynamic performance. The pump power should be chosen 1-2 dBs above its required open-loop value, with all channels present, for the required signal gain. The laser wavelength should be placed either close to the unity-gain region of the clamped gain profile, or at its peak.

Wavelength conversion has been shown to reduce the probability of blocking in both circuit-switching and packet-switching wavelength routed optical networks (WRONs). The effectiveness of the blocking reduction depends on the topology, and is known to be best for meshed topologies, where the average number of hops per path is large. This paper shows that by exploiting wavelength conversion, routing without buffers, known as hot-potato, becomes an interesting option for packet switching WRONs with meshed topologies, such as Manhattan Street (MS) Network and ShuffleNet (SN). The results show that, by using more than 4 wavelengths, a 64 node MS or SN network can work at full load with a hop delay within one hop from its lowest achievable value. We also show that using delay-line routing buffers at the node is a much more effective way of reducing blocking than using wavelength conversion.

Sun et al. (see Electron. Lett., vol.32, p.1490-91, Aug. 1996) reduced the set of coupled first order nonlinear partial differential equations determining the wavelength-dependent, time-varying doped-fiber amplifier gain into a single ordinary differential equation (ODE). In this paper we further simplify the ODE, greatly enhancing its utility as an analysis and design tool. We find that the gain dynamics are completely specified by the total number of excited ions. We demonstrate that channel addition causes much faster transients than channel dropping in wavelength division multiplexing networks. We approximate the solution of the ODE by an exponential with the time constant given as a function of the amplifier parameters.

The teletraffic/transmission performance of regular two-connected multi-hop cell-switching optical networks in uniform traffic under hybrid store-and-forward is presented. The Manhattan street (MS) network and Shufflenet (SN) are compared in terms of average queueing delay, queue size, transit delay, throughput, and bit error rate (BER) for intensity modulation/direct-detection (IM/DD) both analytically and by simulation. A hybrid semi-transparent store-and-forward node architecture is presented. Cells are electronically stored just in the case of conflict to avoid deflection, otherwise the cells will traverse the node without opto-electronic conversion. This architecture performs well, in terms of throughput, propagation delay and BER. It is also shown that by combining deflection routing with the store-and-forward scheme the network can accommodate two different bit-rates. This suggests that the proposed hybrid scheme may have good potential for future multimedia networks.

The transmission performance of regular two-connected multi-hop transparent optical networks in uniform traffic under hot-potato, single-buffer deflection routing schemes is presented. The Manhattan Street (MS) network and ShuffleNet (SN) are compared in terms of bit error rate (BER) and packet error rate (PER) both analytically and by simulation. We implement a novel strategy of analysis, in which the transmission performance evaluation is linked to the traffic randomness of the networks. Amplifier spontaneous emission (ASE) noise, and device-induced crosstalk severely limit the characteristics of the network, such as propagation distance, sustainable traffic, and bit-rate. Our results indicate that under the same load the BER performance of single-buffer deflection routing is worse than hot-potato. However, at BER=10/sup -9/ single-buffer has a higher throughput than hot-potato. It is shown that the feasibility of deflection routing in transparent networks with MS and SN topologies heavily depends on the power coupling coefficient of the routing space switch used in each node.

This paper presents an extension of a known analytical model for the performance evaluation of nonpriority deflection routing networks in uniform traffic. The extension allows the analysis of improved access techniques. The key features of the analytical technique are described by casting it in a very simple setting: nonpriority hot-potato in a two-connected slotted shufflenet (SN) network. Results are presented for three access techniques: transmit-no-hold (TXNH), transmit-hold (TXH), and bypass queuing (BQ).

The steady state behavior of regular two-connected multihop networks in uniform traffic under a hot-potato and a simple single-buffer deflection routing technique is analyzed for very high bit rate optical applications. Manhattan Street Network and ShuffleNet are compared in terms of throughput, delay, deflection probability, and hop distribution both analytically and by simulation. It is analytically verified that this single-buffer deflection routing technique recovers in both networks more than 60% of the throughput loss of hot-potato with respect to store-and-forward when packets are generated with independent destinations. This gain, however, decreases to below 40% when the average message length exceeds 20 packets.>

New single-receiver/single-transmitter/single node structures for two-connected multihop transparent optical packet-switching networks with deflection routing are introduced. A ShuffleNet topology in uniform traffic is adopted to compare various shared optical memory schemes and their control algorithms. These simple structures minimize the number of crossbar switches needed or the node and have moderate control complexity, while still yielding high throughput and low delay. Analytical results are obtained by an extension of the existing theory and verified by simulation.>

The steady-state behavior of regular two-connected multihop networks in homogeneous load under hot-potato and single-buffer deflection routing is analyzed for ultra-fast optical applications. The Manhattan street network and ShuffleNet are compared in terms of throughput, delay, and deflection probability both analytically and by simulation. It is analytically verified that single-buffer deflection routing recovers in both networks more than 60% of the throughput loss of hot-potato with respect to store-and-forward when packets are generated with independent destinations. This gain, however, decreases to below 40% when the average message length exceeds 20 packets.>