>Fergeddabout SCMR, it's already old news. ;-))
  if you'll excuse the intrusion, ray, on that note, i normally wouldn't expect to hear a pure optical play and a SONET integrated chip maker both mentioned positively in the same breath (see the italicized paragraph below in this Wirbel piece).
  with all the NFOEC, Telecosm 99 and Geneva photonic news lasered about, i'm not so sure this should make me less nervous about holding a small basket of comm IC players, say longer than crack-cocaine possession conviction (in DC).
  care to enlighten me here (bad pun unintended)?
  thanks for your response earlier today; i owe you a PM.
  -chris.
  -----
  Electronic Engineering Times October 04, 1999, Issue: 1081 Section: News Big players, startups push fresh fiber concepts -- Sonet shunted aside as packets go optical Loring Wirbel techweb.com
  CHICAGO - The central role of Sonet as the optical backbone of today's networks is rapidly coming to an inglorious end, according to developers at the National Fiber Optic Engineers Conference here last week. But from the remains of the Synchronous Optical Network and the rise of wavelength-based broadband services, new concepts for building an optical infrastructure more closely allied to packet data are emerging from both major players and more than 25 startups.
  Many of the new companies are exploring the transition wherein passive dense wave division multiplexing (DWDM) is being supplanted by optical transponder equipment that performs add-drop multiplexing (ADM) and cross-connect in the optical domain, even though true optical switching is months, if not years, from viability. Because decisions made in the design of transitional optical equipment could affect the fate of future all-optical switches, the debate over Sonet's role has become critical to the trajectory followed by many startups.
  Amid that uncertainty, a surprising consensus emerged here that Sonet's role as a dedicated layer must end, with its functions split between the optical layer and the network Internet Protocol (IP) layer. Sonet is shifting to an embedded role not because it exhibits any limitations for switch protection or performance monitoring, but because the fiber-based standard was optimized for circuit -- switched voice -- and today, backbone networks are overwhelmingly being driven by packet data.
  "Use Sonet as a framing protocol, not as a layer in the network," said Anthony Wright, a member of the senior technical staff at Williams Communications Inc. (Tulsa, Okla.). "Sonet as a standalone piece of equipment, as ADM or as cross-connect, has become obsolete."
  There are many concepts for preserving simplified functions of the Sonet payload envelope at the Internet Protocol layer, including "digital wrappers," promoted by Lucent Technologies Inc., and a moribund effort by the Optical Interoperability Forum to define a Sonet Lites spec. Performance monitoring, meanwhile, can be driven down to the optical layer in any piece of equipment that preserves some form of optoelectronic conversion, in essence including Sonet byte monitoring whenever an optical path is switched electronically.
  That's good news both for such new-breed optical routing companies as Monterey Networks Inc. (Richardson, Texas; recently acquired by Cisco), as well as such chip vendors as Vitesse Semiconductor Inc. (Camarillo, Calif.), which can manufacture devices for Sonet-like channel aggregation inside an optical ADM without having to meet all of Sonet's jitter specs.
  But both Monterey chief technology officer Bill Szeto and Vitesse director of telecom product marketing Greg Borodaty offered similar caveats about embedding Sonet performance monitoring in Layer 1 hardware: Carriers that grow accustomed to preserving some Sonet functions in the physical-layer hardware of first-generation optical equipment may be reluctant to leap to true optical switches in the future, if doing so would mean abandoning the suite of Sonet's restoration and regeneration features.
  Lucent's concept for adding a digital wrapper around an optical-channel payload is among the most advanced concepts in the Sonet replacement field. Paul Bonenfant of Lucent's Holmdel Labs explained that the wrapper adds network management and optional forward error correction functions to any type of packet data, including asynchronous transfer mode, Simplified Data Link, Internet Protocol and Gigabit Ethernet. The wrappers allow the most efficient transition for networks performing direct IP mapping over WDM, Bonenfant said.
  But Eugene Park, director of strategic marketing at optical newcomer LuxN Inc. (Sunnyvale, Calif.), argued that "digital wrappers are just Sonet all over again." LuxN instead proposes ColorSIM-signal-integrity monitoring based strictly on wavelength management, in which signal reshaping, regeneration and retiming are carried out solely within the optical layer.
  In the optical transponder product LuxN plans to introduce by year's end, embedded monitoring will combine with a software-controlled remote provisioning tool, called ColorValve, and with the ability to manage and monitor signaling nonintrusively on the same wavelength as that carrying the data. LuxN's plans assume a fairly proliferated fiber buildout based on dark fiber, since its architecture calls for optical managers in the central office and optical-access managers at customer sites.
  Simpler mapping
  Donald O'Connor, section leader at Fujitsu Network Communications Inc. (Richardson, Texas), said the advent of IP label-swapping methods such as multiprotocol label switching has enabled Sonet to simplify functions in mapping packetized data. Fujitsu has proposed two versions of Sonet Lite: one for metro services and one for long distance. The bulk of section overhead within Sonet would be taken out of the metro version of the protocol; in the long-distance version, all of the section overhead is retained except for J0/Z0 bytes, which are not needed for packet transport.
  The Lite effort is still at an early conceptual stage, O'Connor said, and has not been formally proposed to the Optical Interoperability Forum or similar bodies.
  G. Emory Anderson, who analyzes customer WDM systems at Telcordia Technologies Inc. (formerly Bellcore; Morristown, N.J.) said that new-breed optical ADM and optical cross-connect platforms already are embedding section parity and framing bytes from Sonet into their optical products.
  "A lot of DWDM systems are absorbing many of Sonet's functions already," Anderson said. "Even if DWDM systems can't provide synchronization, IP routers in the network can. Sonet was optimized for voice, and its granularity simply does not fit IP very well."
  Williams Communications has been trying out these concepts in a network, linking sites in Alabama and Mississippi, in which ATM switches from Lucent and Ascend are interfaced with DWDM transponder systems from Sycamore Networks Inc. ATM traffic can move directly to DWDM systems at 2.5 Gbits/second, with no intervening Sonet ADM equipment.
  The advances in tiered optical access systems that take on Sonet functions has split the optical access market into four distinct realms, said Hugh Martin, president of Optical Networks Inc. (San Jose, Calif.): long-haul DWDM, optical cross-connects, metropolitan core rings and metropolitan access rings. Most of the market's new entrants are aiming at metropolitan access; a partial list includes Alidian Networks, Astral Point, Amber, Cyras, Siara Systems, Mayan Networks, Quantum Bridge and Luminous Nets.
  System and component vendors alike presented new concepts for optical transport on the show floor and in technical sessions at last week's conference.
  Kestrel Solutions Inc. (Mountain View, Calif.), a heavily funded startup with roots in Lockheed and Lucent, partially unveiled its strategy and showed off prototypes of its TalonMX optical ADM system. Kestrel chief executive Michael Rowan said the company's founders had rejected both TDM and WDM as inappropriate for metropolitan fiber access, instead looking to frequency-division multiplexing (FDM) as a base for effectively carrying 155-Mbit streams on fiber.
  He noted that Kestrel's founders have a history with FDM, having worked on development projects employing the approach for space-based broadcast networks. Kestrel, he said, has been intent on leveraging modulation subsystems that would reuse design elements from space-access systems.
  The startup applies its approach of using existing sub-systems to modulation and electro-optical conversion. Standard DSPs perform modulation, equalization and error correction in each channel. Individual channels are stacked by frequency and a single laser and photodetector perform all E/O conversion. A 10-Gbit/s stream can be carried in 20 GHz of bandwidth, Kestrel said.
  Dawn Hogh, the startup's vice president of marketing, said a primary goal in designing TalonMX was to ensure that optical access equipment would cause only minimal disruption to the way legacy Sonet and ATM traffic is handled. Many carriers are slower to deploy new transport equipment than data-oriented OEMs assume, she said, so developers pushing the edge of DWDM can fail to look at optimal ways of carrying existing traffic.
  Startup Chorum Technologies Inc. (Richardson) came to the conference with a suite of active optical switching elements dubbed PolarWave, that are based on mature liquid-crystal technology. It also rolled passive interleaved filters based on polarization control technology. They can be layered in channels spaced at 25 GHz or narrower, allowing extremely dense multiplexing equipment to be built.
  Optical processors
  Chorum will combine the two technologies with the use of DSPs for optical signal conditioning in what it calls optical processor modules, but the company said it does not intend to move into system sales.
  Chorum is bound to get some near-term competition in liquid-crystal switches from SpectraSwitch Inc., a Santa Rosa, Calif., startup that plans to offer switching matrices and WDM filters early next year.
  Meanwhile, a team from Templex Technology Inc. (Eugene, Ore.), presented its first public paper on a concept called lightwave code-division multiple access. Borrowing CDMA concepts from radio networks, the technique uses fiber gratings as coders and decoders, superimposing multiple encoded data streams on wavelength channels.
  SilkRoad Inc. (San Diego), which raised eyebrows last fall in claiming single-wavelength systems that could replace WDM, revealed more of its "refractive synchronization" concepts in a conference paper. SilkRoad uses Mach-Zehnder modulators as "optical tweezers," tapping focused laser light to create a field gradient that causes photons to spiral toward the beam with a specific angular momentum. Photons travel in specified Laguerre orders, and the highest can be used as carrier signals.
  SilkRoad engineering director Thomas Myers said a patented form of distributed feedback implements refractive-synchronization-based very high-speed bidirectional communications. |