For multiuser, local, and metropolitan area networks, lightwave technology provides an opportunity to share an enormous network capacity among all network users. However, the selection of an appropriate media access control algorithm may be more important than the fiber bandwidth and the transmitter power in determining the network throughput. We compare a variety of architectures and access protocols for multi-channel packet networks. Included is a recent architecture that uses coherent optics to address the media access issue.

We propose a new broadband local area network, STARNET, based on a physical passive star topology. STARNET implements both a packet network and a high-speed WDM circuit interconnect, which are simultaneously available to every node of the network and operate independently. As a result, STARNET supports very diverse types of traffic in an optimal way. Each node requires only two lasers and its structure permits to effectively achieve frequency stabilization for the whole network. An effort toward an experimental demonstration of a 4-node, 3 Gbit/s per node, FDDI-compatible (at the packet network level) STARNET is currently in progress at Optical Communication Research Laboratory, Stanford University.

A lightwave network consisting of geographically distributed access stations that communicate over a fiber infrastructure is considered. The access station electro-optic transceivers tap directly into the optical medium and transmit/receive on different wavelengths, thereby creating wavelength multiplexed channels that form a logical connection diagram among the stations. Such lightwave networks have the ability to superimpose a variety of logical connection diagrams, or virtual topologies, on top of the underlying optical infrastructure by proper assignment of transmit and receive wavelengths to the network stations. Furthermore, the use of transmitters and receivers that are tunable over a large portion of the optical band allows the network to be dynamically reconfigured. The connection diagram could thus be updated in response to changing traffic patterns and failure/recovery of access station equipment. Potential applications of rearrangeable lightwave networks range from a distributed ATM switch with traffic changing continuously as calls are established and terminated, to metropolitan or wide area networks with slowly changing traffic patterns. In the distributed switch, reconfiguration may occur frequently, and a new connection diagram would be sought each time a new call cannot be accommodated by the switch and would otherwise be rejected. In the wide area network setting, reconfiguration would take place less frequently as the expected traffic pattern slowly and predictably changes over time. This paper describes practical approaches for the reconfiguration phase, during which the network connection diagram is changed from an initial to a final configuration.

TeraNet is an experimental optical network which offers user access rates as high as one Gigabit per second (Gbps). The network provides both circuit switching and ATM packet switching in a hierarchical fashion. Circuit switching is accomplished through the use of tunable optical WDMA/Subcarrier-FDMA channels. An overlaid multihop technique that uses a sub-set of the underlying optical circuits, constructs a packet switching environment. Circuit switched services interface to the passive optical 'ether' through Media Interface Units (MIU). A 3 X 3 switching node called Network Interface Unit (NIU), with 1 Gbps capacity per port, is used as an ATM access port to the packet switching layer. The ATM overlaid network incorporates a traffic control architecture that supports multiple traffic classes. Output buffers and bandwidth are shared according to a resource allocation concept called Asynchronous Time Sharing (ATS).

A crucial component in a future teraflop, distributed, massively-parallel processing system is an interconnection network capable of reaching an aggregate capacity of a terabyte-per-second with gigabyte-per-second access rates for each user. This is the target of the architecture described here, which is the optoelectronic extension of the interconnect previously used in a shared-memory, distributed, MIMD supercomputer. Because of its inherent flexibility, at a lower performance level this same architecture could also meet the growing near-term interconnect needs within and between current large computing centers. The architecture provides small block transfers, a source-destination latency approaching the physical minimum, high average and burst user bandwidth, and expandability to thousands of distributed users. The packet payloads remain in optical format from source to destination while being switched in the multi-state, multi-path network. Large-scale data compression by wavelength results in packets of a few nanoseconds length. Self-routing at intermediate nodes is achieved with brief excursions of the control headers into the electronic domain, where pipelined processors implement a hot potato/deflection switching protocol. A system demonstration is currently under construction.

Simple topologies are proposed for linear lightwave networks (LLN). Tree-based networks which preclude the possibility of certain violations are compared to networks that allow violations to occur but achieve better load distribution within the network. The performance of each topology, from the point of view of blocking probability, is compared by a simulator using a call routing algorithm based on the K-shortest path algorithm. It is found that general topologies that allow for better load distribution perform better than tree-based topologies.

In this paper we study the synchronization problem that exists in the linear lightwave network (LLN), a recently proposed fiber-optic architecture. The currently proposed solution prevents efficient implementation of time division multiple access (TDMA) schemes for sharing the common broadcast medium. The problem is typical of passive optical networks with a tree physical topology. In this paper we propose to overcome the synchronization problem in LLN by properly selecting the routing of the optical signals. The proposed scheme provides efficient utilization of the multiaccess channel with relatively low synchronization complexity.

The recent demonstration of ultrafast, cascadable, all-optical soliton gates, although with long latency and at an early stage of research, opens the possibility of niche exploitation in architectures whose performance is primarily limited by the absence of a few such logic elements. A candidate system is a widely distributed, self-routing short packet, slotted ring system running at peak rates well beyond that of the conventional electronic hosts at each access node. We describe here an architecture for a system with a 1.25 GHz packet rate, 32- bit payload, and 100 GB/s peak bit rate serving a few hundred user nodes. Optical format is retained by through-going node traffic, so that the overhead of conversion to/from electronics is incurred only at the source and destination. This design effort has served to sharpen our understanding of the strengths and weaknesses of using such gates in carefully chosen applications.

Polarization modulated coherent optical communication systems are quite interesting for their simplicity, low phase noise induced penalty, and insensitivity to state of polarization fluctuations if a suitable tracking is adopted. In this paper different kinds of binary and multilevel polarization modulated systems are presented based on Stokes parameters detection. The structure of this class of systems is described and the polarization tracking algorithms are considered. Quantum limit performances are evaluated and the phase noise induced penalty is obtained both in the presence of a sample and hold decision device and in the case of postdetection filtering.

We propose and analyze a novel binary coherent optical transmission scheme based on continuous polarization modulation. Its bandwidth efficiency is better than that of both PSK and MSK. The signal spectrum is independent of the polarization transformations along the fiber. Its phase-noise tolerance is similar to FSK and its sensitivity is 3 dB better than that of single-filter FSK.

Digital switching of the state of polarization (SOP) has been proposed in optical coherent communications for either polarization-independent or SOP-coded transmissions. In both cases external SOP modulators with high insertion losses have been employed. To avoid this penalty, the use of a twin-laser module synchronously operating both in polarization and frequency shift modulation conditions is presented. The feasibility of the relevant schemes for both the aforesaid system types is demonstrated through numerical simulations at 1.55 micrometers wavelength and 155 Mbit/s bit rate.

The n-ary pulse position modulation is a means by which excess optical fiber bandwidth can be exchanged for improved receiver sensitivity. Here, we consider both direct detection (PIN- BJT and PIN-FET) and coherent n-ary PPM over a range of fiber bandwidths and PPM word sizes. Our results illustrate that n-ary PPM offers an improvement of between 5 - 11 dB over equivalent PCM systems. Assuming an optical fiber attenuation of 0.2 dB/km this represents an increase in regenerator spacing of between 25 - 55 km. The theoretical results demonstrate that n-ary PPM is a promising modulation format and may have potential for future telecommunication routes.

In this paper, the performance of a direct detection optical DPSK systems using a dual detector receiver is analyzed. The filtering effects on the intensity noise, resulting from the interferometric conversion of laser phase noise, are also investigated. Linewidth requirements are estimated and compared with that of heterodyne DPSK systems. It is show that linewidth requirements can be relaxed by a factor of about 0.58, when filtering effects are taken into account.

This paper analyzes an optical wavelength division multiplexing system (WDM) with subcarrier multiplexing (SCM). The pump laser is tuned to amplify the corresponding optical carrier by fiber Brillouin amplification (FBA) in WDM for the desired group of SCM signals and then a microwave tuner is used to select the desired channel in this group. This system has the benefits of eliminating the need of polarization control, the ability of phase noise cancelling due to the 'squaring' photodetection process of the selected optical carrier together with its SCM channels, and enhancement of optical receiver sensitivities by amplification of the carrier.

We analyze the performance of on-off-keying modulated signals in the presence of phase noise for different envelope detection structures. We obtain closed-form expressions for the error probability conditioned on a random envelope, and remove the conditioning via approximations. We also provide a tight lower bound in closed-form. We also compare the performance with that of frequency shift keying.

A simple technique for analyzing the sensitivity and bit-error rate performance of direct- detection lightwave receivers using optical preamplifiers is presented. Our analysis provides closed-form expressions for the system performance, and includes the impact of phase noise. For zero (negligible) linewidth, our theory predicts an average signal energy of 42.3 photons/bit at BER equals 10-9. In comparison, at BER equals 10-9 Ref. 1 predicts an average signal energy of 38 photons/bit.

Direct-detection (DD) optical fiber frequency division multiple access (FDMA) is a simple and practical alternative to optical heterodyne FDMA. Previous works on estimating the performance of DD optical FDMA networks have usually focused only on the (linear) crosstalk degradation, typically relying on simplifying approximations such as the dominance of the adjacent channel interference. Other sources of performance degradation such as signal loss due to optical (predetection) and electrical (postdetection) filtering, intersymbol interference (ISI) due to the optical filtering, and channel beats (or nonlinear crosstalk) are usually ignored. This paper presents a more precise analysis that takes into account the effect of optical and electrical filtering, ISI, and linear crosstalk. The model used here is valid for arbitrary optical filter transfer functions and received pulse shapes. We consider in this paper only On-Off-Keyed (OOK) modulation; extending the analysis to Frequency-Shift-Keyed (FSK) modulation is straightforward, but is not discussed here.

A study of the multichannel coherent CPFSK communication system employing Reed-Solomon codes is reported. The crosstalk between adjacent channels, the impact of the laser phase noise, and selection of the optimum code rate are investigated. By applying a suitable coding scheme, the laser linewidth can be greatly relaxed, and the receiver sensitivity will be remarkably improved. A 100 Mb/s system is used to verify the coding benefit. The results show that the impact of the laser linewidth can be largely alleviated by applying a (255, 223) Reed-Solomon code. When the linewidth is fixed ((Delta) (upsilon) T equals 2%), this code can improve the receiver sensitivity by more than 15 dB.

The first 8 Gb/s coherent optical detection system using two closely spaced subcarriers with 4 Gb/s QPSK on each subcarrier while maintaining an RF bandwidth efficiency of 1 bit/sec/Hz is reported. The receiver sensitivity is -26.5 dBm, with less than 0.5 dB penalty due to the other channel. It is expected that these techniques can be extended to transmit 16 Gb/s while using only 2 Gb/s baseband electronics.

We investigate the effect of polarization mode dispersion (PMD) on a coherent optical distribution system with shared local oscillator (LO). The PMD causes mismatch in the states of polarization (SOP's) of the message and LO signals. Because coherent reception relies on the signal SOP, the intermediate frequency signal is degraded by the PMD. Simple expression is derived to estimate the performance degradation.

We present exact solitary wave solutions for multigigabit systems operating in the femtosecond time domain. Our results show different behavior from the traditional nonlinear Schroedinger equation.

The application of a novel integrated LiNbO3 device, combining the functions of an electro-optical phase modulator as an acousto-optical TE/TM mode converter, in a 34 MBit/s DPSK system is described. The device it suitable for use as passive remote transmitter unit in coherent single-laser communication systems based on the self-heterodyne concept.

The positive effect of widening the 1.5 micrometers Erbium fluorescence band provided by alumina on germanosilicate core glass of active fibers is partially obtained also in magnesia silica host, but to a lesser extent, while zirconia silica fibers do not exhibit modification of the emission band. Excited state lifetimes, absorption, shape, and width of the band in different glass hosts were measured with the aim to optimize the core glass composition.

We describe experimental results of frequency up-conversion spectroscopy in Pr3+ doped monomode silica based fibers. Visible to ultraviolet and near-infrared to visible frequency conversion is demonstrated and analyzed using nanosecond and picosecond excitation sources.

Using a 1GHz, 1W microwave source to drive a comb generator con— sisting of inicrostrip circuit and a SRD, We produced electrical pulses, with a pulse width of 70-l2Ops and an amplitude of l5V, which, along with 3OmA DC bias, were applied directly through a microstrip matching circuit to a l.532pm InGaAsIn/InP DFB-LD, which threshold current is 2Oma. the optical pulses of less than lOOps were produced.