Soliton transmission, optical crossconnects advance optical networking
By STEPHEN HARDY
The post-deadline papers at the Conference on Optical Fiber Communication (ofc) each year provide attendees with a glimpse of what`s new "in the lab"--as well as the opportunity to handicap which technologies will find their way out of research and development and into the field. This year`s conference, held February 23 to 27 in San Jose, CA, proved no exception. While wavelength-division multiplexing (wdm) dominated the proceedings in the post-deadline presentation rooms as clearly as it did everywhere else in the conference, two other networking technologies--soliton transmission and optical crossconnects--showed signs of breaking into the applications arena in the near future.
Soliton transmission seems by far the more esoteric of the two emerging technologies. As reported in Lightwave last month (see Lightwave, February 1998, page 100), solitons are a type of optical pulse that resists dispersion and distortion over great distances. While the phenomenon was discovered as far back as 1834, and first investigated in the lab for optical transmission more than 20 years ago, the potential of solitons to increase the signal capacity of fiber has so far successfully eluded capture, despite work in the area worldwide.
However, Pirelli Cables and Systems S.p.A., Milan, Italy, appears to be close to commercializing such a system. The company teamed with mci Telecom- munications, Richardson, TX, to conduct a field trial of a 4 ¥ OC-192 (10-Gbit/sec) dispersion-managed soliton system that successfully operated bidirectionally over 450 km of standard fiber, and unidirectionally over 900 km. The transmission occurred in the 1550-nm erbium band, and the fiber was selected to have high polarization mode dispersion coefficients.
The figure on page 20 shows the initial route configuration. The soliton pulse generators were placed after the commercial OC-192 non-return to zero (nrz) transmitters; the soliton return to zero-to nr¥converters were placed before the commercial nr¥receivers. While previous experiments using commercial nr¥OC-192 systems had required the use of additional optoelectronic regenerators at the mid-route site, the latest demonstration did not. Thus, capacity on standard fiber routes of this length could be increased merely by adding additional OC-192 equipment at the terminal sites.
The average system dispersion was reduced to 0.75 psec/nm-km from west to east (on 1549- and 1557-nm transmitted wavelengths) and to 1.85 psec/nm-km in the opposite direction (1553- and 1541-nm transmission wavelengths) using a combination of multiple-fiber Bragg gratings (dispersion-compensating gratings--dcgs) and dispersion-compensating fiber (dcf). The dcg/dcf devices resided at each amplifier site. This gave two different dispersion maps for the two directions. Interferometric filters with 0.5-nm bandwidth demultiplexed the received signals at the terminal sites. The mean differential group delay for the route was 9.2 psec.
To test transmission at a longer distance, the 1541-nm signal was removed from the east-west direction. This boosted the signal-to-noise ratio on the 1553-nm channel by 2 dB. This channel was looped back on a second fiber for a total system length of 900 km. The combination of five multiple dcgs and seven dcf units reduced the average chromatic dispersion of the network to 1.15 psec/nm-km. The mean differential group delay was 26.9 psec.
According to Pirelli sources at the company`s exhibit, the equipment used in the demonstration was a specially adapted version of the company`s bds 4 ¥ 4 system. Having demonstrated four-wavelength soliton transmission that provided a twofold increase over installed OC-192 nr¥systems, the sources indicated that the company would launch a commercial soliton transmission product in the near future--perhaps as early as the second quarter of this year.
Optical crossconnects
For their part, optical crossconnect systems have already been announced as commercial products--although some in the industry have questioned whether the subjects of these announcements were more fantasy than fact. The naysayers began to hold their tongues when Astarté Fiber Networks Inc., Boulder, CO, announced recently that WorldCom had been conducting field trials of its Model 7250 StarSwitch optical crossconnects. The two companies used the ofc post-deadline sessions to provide the first public details of the implementation. The crossconnect was a multimode system that accepted 72 transmit/receive pairs, provided 100-msec switching, was remotely controllable, suffered less than -50 dB of crosstalk, and had an insertion loss of 2.5 dB or less.
The multimode switch, according to the presenter, Marvin Young of WorldCom in Tulsa, OK, was the only product available on the market at the time WorldCom was investigating the technology. Because WorldCom used singlemode OC-12 fiber transmitters and receivers on its network, the company first had to test the compatibility of the crossconnect with the existing network. It concluded that the additional loss induced by the multimode to singlemode mismatch was well within system margins, and that dispersion and modal noise would not be a problem.
The company purchased 24 crossconnects, which were fully deployed by January 1997. The systems were tasked to manage OC-12 (622-Mbit/sec) transmission within the network sites--a task they continue to perform. Each site contains a pair of crossconnects. One manages work circuits, the other provides protection circuits. Redundant industrial-environment computers, which run a network management program developed by WorldCom, provide system control. They can be accessed from the company`s network control center in Tulsa.
The systems have been running for a year, and Young said that WorldCom is satisfied with their operation.
Of course, the United States is not the only area of the world experimenting with optical crossconnects and other forms of optical network technology. For example, the European Commission has funded the Advanced Communications Technologies Services Optical Pan-European Network (acts open) project. During the ofc post-deadline sessions, project participants described the first of two cross-border field trials of wdm networks tied together with optical crossconnect systems.
The trial took place in Norway and Denmark. The network consisted of a 4 ¥ 2.5-Gbit/sec wdm terminal in Oslo, Norway, linked by 350 km of G.652 fiber to the Norwegian town of Arendal, where the optical crossconnect was housed. The crossconnect was a subequipped 4 ¥ 4, four-wavelength, 2.5-Gbit/sec all-optical system. It was based on a "broadcast-and-select" architecture and used semiconductor optical amplifier technology, such as clamped-gain integrated optical gate arrays and all-optical 2R-regenerating interferometric wavelength converters. The facility also had 2 ¥ 2 transit connectivity and add/drop mechanisms.
From Arendal, the network extended via two routes to Thisted in Denmark. A western route crossed the North Sea from Kristiansand in Norway to Thisted; the 140-km of submarine cable contained dispersion-shifted G.653 fiber. (A return path from Kristiansand to Arendal also was installed to allow up to three optical crossconnects to be cascaded in a loop.) Meanwhile, an eastern route linked Arendal to Hjorring in Denmark via another 140-km submarine cable using G.653 fiber. In Hjorring, a wavelength add/drop multiplexer provided a partial means of measurement. Hjorring was linked to Thisted via 180 km of G.652 fiber.
In Thisted, a receiver terminal provided full measurements on the optical signal. Transmission equipment included 20 flat-gain fluoride-based line amplifiers that provided 22- to 27-dB gain, +15 dBm typical output power, and a typical noise figure of less than 7 dB. The network also used 80 km of dispersion-compensating fiber.
Technicians put the crossconnects through a series of tests both before and during implementation. The pretests examined the systems` "cascadability"; they revealed bit error ratios lower than 10-12 independent of the sequence length and sensitivity excursions lower than 2 dB for all configurations. Six configurations have been tested in the field trials, including transmission to 1000 km and the cascade of three crossconnects. Q factor assessments at Thisted ranged from 9.7 to 12. Actual measurements of bit error produced results below 10-11.
Both field trials reveal that optical crossconnect technology may be ready for implementation in the very near future. Soliton transmission systems are somewhat farther from deployment--but not as far as may have been assumed. q