While all-optical networking had its origins in the research community a quarter of a century ago, the realization of the vision has not had a straight trajectory. The original goal of the all-optical network was based on keeping the data signals entirely in the optical domain from source to destination to eliminate the so-called electronic bottleneck, and to allow arbitrary signal formats, bitrates, and protocols to be transported. The latter property is referred to as transparency. When all-optical networks were finally commercialized around the turn of the century, however, a modified reality emerged; the quest for transparency was replaced by the more pragmatic objective of reducing the network cost and energy consumption. Moreover, especially for networks of large geographical extent, electronics were still present at some (relatively few) points along the data path, for signal regeneration and traffic grooming. This modified vision captures the state of today's networks, though terms like all-optical and transparent are still used to describe this technology. However, continued advancements are bringing back some aspects of the original transparency vision. In this paper, we review the evolution of all-optical networking, from the early vision to its present vibrant state, which was made possible by great advances in optical transmission and all-optical switching technologies. We describe the numerous benefits afforded by the technology, and its relative merits and drawbacks compared to competing technologies, sometimes referred to as opaque. We also discuss the remaining challenges and future directions of all-optical networking. While all-optical solutions permeate today's access, metro, and core networks, this paper focuses on the core.

With the expected explosive growth in data traffic, the required processing demands at network nodes, e.g., in routers, will likely present a significant bottleneck. Optical bypass potentially alleviates this bottleneck by allowing a wavelength to bypass a node if it contains no traffic that needs to be processed there. We explore the value of optical bypass in gigabit networks by measuring the fractional decrease in required router size one can expect. We consider both ring and mesh topologies, and both uniform and distance-dependent traffic. We show that the reduction in router size can be very significant, with the fractional savings increasing as the size of the network increases and the amount of traffic increases, indicating that this technique will scale well in future networks.

Summary form only given, as follows. In December 1994, a group of major US telecommunications industry players (AT & T, Bell Atlantic, Bellcore, BellSouth, Lucent Technologies, Pacific Telesis and Southwestern Bell) formed the Multiwavelength Optical Networking (MONET) Consortium, with partial support from the US Defense Advanced Research Projects Agency (DARPA), to investigate and advance the vision of national-scale optical networking. This vision is based on the creation of a configurable, reliable, high-capacity optical backbone network layer capable of supporting virtually all current and future telecommunications bit-rate, format and protocol standards across long-distance, local-exchange, and private networks. This talk summarizes the motivation and the status of the MONET project.

We describe some of the results of the Advanced Research Projects Agency (ARPA) sponsored Consortium on Wideband All-Optical Networks in developing architectures, technology components, and applications for the realization of scaleable, wideband, and transparent optical wavelength-division multiplexing (WDM) networks. Our architecture addresses all-optical transport over the wide, metropolitan, and local areas. It utilizes wavelength partitioning, routing, and active multiwavelength cross-connect switches to achieve a network that is scaleable in the number of users, data rates, and geographic span. The network supports two services which can be point-to-multipoint or multipoint-to-multipoint simplex or duplex connections. The A service is a transparent physically circuit-switched service and the B-service is a scheduled time-slotted circuit which is transparent within its time slots. We have developed a 20-channel local and metropolitan area WDM testbed deployed in the Boston area, now undergoing characterization and experimental applications.

The general principles, implementation, and performance of a 1-Mb/s time-division-multiple-access (TDMA), slow frequency hopping and coding, 1.5-GHz radio communications system for a medium-sized office building are discussed. TDMA is provided for service flexibility and slow frequency hopping and coding for immunity against multipath fading and interference. Measurements show that, with a hallway-mounted distributed antenna system installed on one floor of the building and with the mobile unit transmitting only 1 mW of peak RF power from anywhere on that floor, only a single 384-b frame out of a total of 200000 transmitted frames suffered an unrecoverable error. This kind of performance is comparable to that of wired data modems. Similar performance was obtained from a central antenna covering the same floor. However, the required transmitted power in this case was 100 mW. Implementation issues, such as spectrum allocation, cellular subdivisions, and spectrum efficiency, are also discussed.>

In future optical networks on frequency division multiple access (FDMA), e.g. TV distribution, optical amplifiers will be required to assure adequate reception. A key figure of merit is eta , the efficiency with which channels can be packed into a given bandwidth relative to the ideal packing of 100% corresponding to abutting channels (channel spacing=channel bandwidth). Packing can be severely limited by intermodulation distortion due to inherent amplifier nonlinearity. Assuming constant envelope signaling, the authors use a recently developed amplifier model to show that, by employing practical coding methods, eta can be greatly increased. A simple, general algebraic formula for eta is derived in terms of amplifier, receiver, and code parameters. For illustrative purposes the authors consider the packing of 50-Mb/s channels into a 1-THz optical bandwidth, assuming typical amplifier and receiver parameters.>

It is shown that when the amplifier is driven near saturation, its inherent nonlinearity causes significant bit-pattern-dependent pulse distortion, particularly in the bit-rate range between about 2 and 32 GB/s. Without proper countermeasures, this distortion can degrade system performance appreciably due to two basic mechanisms. The first, which can result in a system power penalty of as much as 10 dB, occurs in a standard decision circuit that automatically sets the threshold voltage to the average signal level, rather than in the middle of the eye opening. The second mechanism, which occurs even with the threshold set properly, is due to the nonlinear enhancement of the simple linear intersymbol interference (ISI) within the receiver filter. For example, computations of system performance at 8 Gb/s using an RC filter that gives a quite acceptable 10% of eye closure under linear conditions show that when the amplifier is driven to its saturation output power level, this mechanism causes a system power penalty of about 1 dB, which increases to about 4.5 dB when the power is doubled. Interestingly, with the proper threshold setting, an ideal integrate-and-dump receiver, which introduces no ISI, is shown to suffer no power penalty due to amplifier nonlinearity.>

The authors describe and analyze a flexible, robust, multichannel time-division-multiple-access (TDMA) radio communications system that is ideally suited for portable as well as mobile indoor use. It uses noncoherent signaling and incorporates slow cyclical frequency hopping and burst-error-correcting coding to provide low-error communications at aggregate rates of several megabits per second without the need for antenna diversity or adaptive equalization. The system can be implemented in a small building using a centralized architecture or in a large building using a distributed antenna system. The TDMA protocol allows the various users the flexibility of choosing their own desired rates of data communications. Moreover, the capability of using different levels of coding gives the users the option of selecting their own error performance. The ability to combat interference is one of the salient features of this system.>

We consider a set of uncoordinated users employing an ALOHA protocol to transmit packets to a central base station in a local radio environment, e.g., in an urban area or within a building. Because of the different distances between the various users and the base station, the signals received from different users have substantially different power levels. This near/far phenomenon gives rise to a capture effect at the receiver such that when several packets are transmitted simultaneously, the receiver has a good chance of accurately detecting the packet arriving with the highest energy. Moreover, capture can be made near perfect in this environment through the use of properly design spread-spectrum signaling technique that enables the receiver to differentiate among the arriving packets based on both power differences and differing arrival times. We investigate variations to the conventional ALOHA protocols that take advantage of perfect capture to reduce the delays and increase the throughputs of all users, including the furthest one. Three protocols are introduced and compared, which we call "persistent ALOHA with capture" (PAC), "gated ALOHA with capture" (GAC), and "gated ALOHA with capture/exhausted" (GAC/E). Two goals are considered in the comparison: an equatable delay profile among the users, and a small expected number of simultaneously transmitted packets. The latter is quite relevant for the practical implementation of a perfect-capture receiver. We conclude that the GAC/E protocol is most desirable from the twin points of view of delay profile and receiver implementation.

We have investigated the performance of the times-four method of carrier recovery in a digital radio link transmittingM-level quadrature amplitude modulation (M-QAM), whereMcan be 16, 64, 256, etc.. With the help of a recently built multipath fade simulator, we measured rms phase jitter for various combinations of fade notch frequency and fade depth. This paper reports on the experiment and the associated analysis. The times-four method has the virtues that it does not depend upon accurate data decisions, and it does not interact with the receiver's adaptive equalizer. Our results show that the rms phase jitter using this method, for all but the most severe and improbable fades, could be held to a few degrees. Although the measurements were done only for 16 QAM, the analysis predicts only modest increases for higher numbers of levels. The analysis also predicts, and the measurements verify, that the recovered carrier amplitude shrinks to zero for four distinct combinations of fade notch frequency and fade depth. To our knowledge, this phenomenon has not been reported in the literature heretofore. We show that, from a statistical viewpoint, the damage associated with it (measured in terms of outage time) should be quite small.

The results of indoor multipath propagation measurements using 10 ns, 1.5 GHz, radarlike pulses are presented for a medium-size office building. The observed channel was very slowly time varying, with the delay spread extending over a range up to about 200 ns and rms values of up to about 50 ns. The attenuation varied over a 60 dB dynamic range. A simple statistical multipath model of the indoor radio channel is also presented, which fits our measurements well, and more importantly, appears to be extendable to other buildings. With this model, the received signal rays arrive in clusters. The rays have independent uniform phases, and independent Rayleigh amplitudes with variances that decay exponentially with cluster and ray delays. The clusters, and the rays within the cluster, form Poisson arrival processes with different, but fixed, rates. The clusters are formed by the building superstructure, while the individual rays are formed by objects in the vicinities of the transmitter and the receiver.

The idea of implementing an indoor radio communications system serving an entire building from a single central antenna appears to be an attractive proposition. However, based on various indoor propagation measurements of the signal attenuation and the multipath delay spread, such a centralized approach appears to be limited to small buildings and to narrow-band FDMA-type systems with limited reliability and flexibility. In this paper, we present the results of indoor radio propagation measurements of two signal distribution approaches that improve the picture dramatically. In the first, the building is divided into many small cells, each served from an antenna located in its own center, and with adjacent cells operating in different frequency bands. In the second approach, the building is divided into one or more large cells, each served from a distributed antenna system or a "leaky feeder" that winds its way through the hallways. This approach eliminates the frequency cell handoff problem that is bound to exist in the first approach, while still preserving the dramatic reductions in multipath delay spread and signal attenuation compared to a centralized system. For example, the measurements show that, with either approach, the signal attenuation can be reduced by as much as a few tens of decibels and the rms delay spread becomes limited to 20 to 50 us, even in large buildings. This can make possible the implementation of sophisticated broad-band TDMA-type systems that are flexible, robust, and virtually building-independent.

The "computational cutoff rate," R0, represents a practical measure of the maximum reliable data rate that can be achieved by coding over a given communication channel using a given modulation format, in contrast with the "channel capacity,"C, which represents an idealized theoretical limit on the achievable data rate. Moreover, designing signal sets with good error probabilities using the R0criterion results in a mathematical problem that is much more tractable than that obtained by using the probability of error itself as a criterion. Both of the above reasons establish the importance of R0in communications theory. This paper starts with a brief tutorial background, which reveals the origin and the significance of R0. Next, the problem of achieving R0over the additive white Gaussian noise (AWGN) dispersive or nondispersive channel, using quadrature-amplitude modulation (QAM) with a peakpower constraint, is addressed. The major result is that, for both cases, the optimum transmission signal set is chosen from a discrete distribution. The solution is derived in detail for the peak-power-limited nondispersive channel, where it is shown that the optimum QAM symbols are selected independently from a probability distribution that is uniform in the phase and discrete in the radius. The solution for the corresponding peak-power-limited dispersive channel is obtained only asymptotically, for large signal-to-noise ratio (SNR), where it is shown that the QAM symbols are selected independently from a uniform distribution within a disk in the complex signal space.

In a frequency-division multiple-access (FDMA) satellite communication system, the output power of the TWT (or FET) amplifier is backed from saturation to reduce the intermodulation distortion caused by AM/AM and AM/PM nonlinearities. The required amount of backoff can be considerable, wasting valuable satellite power, unless some scheme is used to compensate for these nonlinearities. A comparative study is presented for the effectiveness of three types of predistortion compensators: (1) a cuber predistortion linearizer, which is the simplest to build; (2) an AM/PM equalizer; and (3) an ideal linearizer, which is quite difficult to realize in practice.

Simple two-parameter formulas are presented for the functions involved in the amplitude-phase and the quadrature nonlinear models of a TWT amplifier, and are shown to fit measured data very well. Also, a closed-form expression is derived for the output signal of a TWT amplifier excited by two phase-modulated carriers, and an expression containing a single integral is given when more than two such earriers are involved. Finally, a frequencydependent quadrature model is proposed whose parameters are obtainable from single-tone measurements.