The plenary talks in that long ago Boston conference sound familiar, if slightly modified by developments in technology and jargon: "Access Network Evolution - Cable vs. Telco"; "Video on Demand: Technologies and Services"; "Digital Video Trunking for Regional Networks"; and "Broadband Services: HFC or ADSL." However, the need to deliver these revenue-generating services was not as acute in 1995, when competition was just on the horizon, as it is now when a flood of wireline and wireless options are appearing.
Attendees at NFOEC 1995 also heard service-related talks about digital crossconnects to support SONET, fibre in the loop, and cable TV initiatives. They learned about new technologies such as optical crossconnects, 10-Gbit/sec transmission over standard fibre, and a multiwavelength reconfigurable all-optical network testbed.
Attendees at NFOEC 2003 heard similar talks about metro/access networking systems and systems management. In particular, attendees were drawn to developments in reconfigurable optical add/drop multiplexers (ROADMs), which carriers such as Verizon are beginning to deploy. These switching/blocking all-optical subsystems have been targeted at long-haul networks, but many companies are developing smaller (four- or eight-channel) ROADMs to provide service opportunities in metro applications.
The most common core technologies for ROADMs are MEMS and liquid crystals (LCs). For example, Active Optical Networks (Fremont, CA) exhibited its MEMS-based ROADM, while Arroyo Optics (Santa Monica, CA) described its LC-based ROADM in private meetings. Corning (Corning, NY) had dominated the LC switching market but dropped out earlier this year. Avanex (Fremont, CA), which bought most of Corning's optical-component division, showed its own LC waveblocker technology. Clarendon Photonics (Newton, MA) took a different tact with its waveguide-based ROADM (see Photo). And in a traditional but proven approach, Meriton Networks (Ottawa, Ontario) said that optical-electrical-optical conversion remains the best approach to ensuring add/drop performance monitoring and network management.
Off the exhibit floor in the technical sessions, researchers from Continuum Photonics (Billerica, MA) described the piezoelectric actuation of their MEMS-based beam-steering switch. Researchers from Movaz Networks (Atlanta) focused on all-digital drives for their MEMS-based switch. And a polymer planar lightwave circuit was the alternative approach described by researchers from DuPont Photonics Technologies (Wilmington, MA). Most ROADMs include a variable optical attenuation function.
A market survey by our sister organisation, KMI Research, supports the hope that ROADMs will actually be deployed in substantial numbers. Metro and access carriers report that the average number of OADMs in a network was 200 in 2003; by 2005, that number is expected to more than double. In long-haul networks, the average number is currently 18 per network; by 2005, the number is seen growing to 188.
KMI believes that ROADMs will play an increasing role in new builds and retrofits, with over 20% of the participants in their study saying ROADMs would constitute 40% of their 2003 deployments of OADMs. By 2005, 31% of the participants expect to deploy more than 40% of their modules as reconfigurable.
Most people at NFOEC 2003 knew that attendance was an act of faith that market conditions will improve. As the past eight years have shown, no one can accurately forecast the future, but the essential need for services and related equipment is predictable.