Mux/Demux Components

Gone are the days when George Gilder would wax lyrical about the wonders of all-optical networks and wave-division multiplexing – and send stocks soaring in component companies like Avanex Corp. (Nasdaq: AVNX).

Now it’s a case of wondering when the market for mux/demux components will revive and which companies (including Gilder's) will still be around when that finally happens (as it will, sooner or later).

It’s a tricky topic, partly because there are so many variables.

For a kickoff, the widgets used to split light into wavelengths and do the opposite (recombine them) have all sorts of applications.

Of course, they’re used in Wavelength Division Multiplexing (WDM) systems, but requirements differ markedly among ultra-long-haul, long-haul, metro, and access equipment. And that’s just for starters. Mux/demux devices are also needed in optical add/drop multiplexers (OADMs), dynamic gain flattening filters, and optical performance monitors.

Some types of mux/demux device also promise to become foundation stones for an almost limitless range of integrated optics developments (see Photonic Integrated Circuits).

Each of these applications requires a different combination of performance characteristics, covering the spacing of wavelengths, losses, the clarity of signals, and so on.

To make matters even more interesting, mux/demux devices can be made using a bunch of different technologies: thin-film filters (TFFs), fiber Bragg gratings (FBGs), arrayed waveguide gratings (AWGs), bulk diffraction gratings, and others. Each one has its pros and cons. As a result, matching applications with technologies is a complicated business – particularly as some technologies have improved markedly in recent years.

The best way of getting a handle on these issues is to begin with the basics. This report aims to help you do this. It starts by reviewing the main technologies and then goes on to identify trends and key players. It finishes up by giving a forecast of how the market might develop.

The author is Roy Rubenstein, senior analyst covering optical components with RHK Inc.

Here is a hyperlinked summary:Technologies
Thin-film filters, fiber Bragg gratings, arrayed waveguide gratings, and bulk diffraction gratings.Trends and Market Players
Mux/demux component trends in long-haul and metro applications, and proponents of the big four technologies in component and system development.Forecast and Prospects
Will the decline continue, and what does the future hold?— Introduction by Craig Williamson, Associate Editor, Light Reading
http://www.lightreading.comThis briefing is based upon a report by RHK: DWDM Multiplexers, Demultiplexers, and Add/Drop Optics

The technology used for mux/demux operations can be divided into two categories, based upon the format in which they are used:

Thin-Film Filters
Thin-film filters are definitely the incumbent technology with the largest revenue share at present, being widely used by system vendors. They represent a well understood and established technology, with the attractive characteristic that their cost scales with channel count, because they're modular in design.

Fiber Bragg Gratings
Fiber Bragg gratings have a smaller but devoted following. FBGs have the advantage of improved isolation performance over TFFs, but give higher losses. They also need a circulator to separate wavelengths, and this adds to the overall cost. Two particular systems vendors are devotees of FBGs: Ciena Corp. (Nasdaq: CIEN) and Corvis Corp. (Nasdaq: CORV).

Arrayed Waveguide Gratings
AWGs are now very much a proven technology but in limited use, as their performance has only recently begun to match that of TFFs. They make use of microelectronic manufacturing techniques: batch processing using 6- and 8-inch silicon wafers. They are a significant mux/demux technology but also an important building block in passive monolithic planar lightwave circuits and hybrid integrated optical devices such as variable optical attenuators (VOAs) and reconfigurable optical add/drop multiplexers (ROADMs).Unlike the first two technologies, AWGs process wavelengths in parallel. However, they require a certain channel count before they can compete on price with technologies such as TFFs – the current crossover point is 16 to 20 channels. An AWG is thus less modular than a TFF device. It also requires careful temperature control and has to battle against the TFF conservatism of carriers and systems vendors. The achievement of AWGs is that they have now largely crossed this barrier to achieve mainstream acceptance.

Bulk Diffraction Gratings
Bulk diffraction gratings are regarded as an emerging alternative to AWGs. Although an old technology, they are relatively new to the mux/demux scene. Advantages include a low cost per port, with relatively simple manufacturing processes.Like AWGs, bulk gratings also have an integration roadmap to create devices such as VOA multiplexers, ROADMs, and dynamic gain equalizers. However, question marks exist over alignment tolerances and the traditionally manual manufacturing approach. Also, bulk diffraction gratings have yet to gain the level of acceptance that AWGs enjoy. Consequently, although many of the main system vendors have trialed the technology, adoption in any significant numbers has yet to happen.

Emerging technologies
There are two further emerging technologies to note: polymer AWGs and Echelle gratings. Echelle gratings are being worked on by MetroPhotonics Inc. (see Echelle Gratings Make a Comeback and Optenia's Loss is MetroPhotonic's Gain). These are not discussed further in this report, as the current market is addressed by the four technologies listed above.

Trends in Long HaulOne trend in the long-haul market for mux/demux components is extra channels in the C-band through more tightly spaced channels: i.e., moving from 100GHz spacing to 50GHz. In predominant use today is 100GHz-spaced 10-Gbit/s signals, with 40-channel AWGs available for use with such systems.Current state-of-the-art technology is 50GHz-spaced 10-Gbit/s wavelengths that use interleavers with 100GHz AWGs or TFFs, rather than true 50GHz-capable mux/demux components. When offering 80 channels in each of the C and L bands this makes a possible total capacity of 1.6 TBit/s. From 2003 there will be a move towards 25GHz spacing at 10 Gbit/s, and later 100GHz spacing of 40-Gbit/s wavelengths, with capacities up to 4 Tbit/s a reality.Trends in MetroMetro networks are characterized by fewer channels and a more relaxed channel spacing of typically 200 GHz. They also require the use of more optical add/drop multiplexers than with long-haul transport. Thin-film filters are still the technology of choice due to cost and granularity advantages associated with lower-channel WDM systems.The introduction of reconfigurable OADMs will drive the demand for novel mux/demux components in both metro and long haul, meaning that nodal applications are a distinct growth area in comparison to applications in the terminal. Optical performance monitoring and gain balancing (combining a variable optical attenuator and a multiplexer to give a VOA-mux) are other nodal ancillary uses.The PlayersThe market for mux/demux components is indeed a crowded one, with more than 60 players in the field. There are over 20 companies working on TFFs, with a similar number working on AWG technologies. Last year saw two of the 20 AWG makers disappear, with Kymata Ltd. being acquired by Alcatel Optronics (Nasdaq: ALAO; Paris: CGO.PA) and Zenastra Photonics failing to obtain new funding (see Alcatel Optronics Acquires Kymata and Zenastra Photonics: RIP). The following is a selection of some of the players addressing each technology, along with some systems vendors that have pet favorites.

RHK Inc. is in the process of updating its forecasts, but this is how things looked at the end of last year:

The WDM optical component market is expected to decline further in 2002. What we are already seeing is that the decline will be towards the pessimistic scenario with a figure of -17% most likely.Optical component vendors are still not offering guidance, and history shows that a rebound is needed in Q1 or Q2 for positive market growth in 2002.This is because the rebound will be at the end of the year, and 2002 will not be flat.The mux/demux bubble certainly burst in 2001, with total market sales expected to drop 60 percent compared to 2000. This market has been hit hard because its main application, the equipping of new fibers, has not been happening. Since carriers are only lighting new light paths, demand for terminal mux/demux equipment has been extremely limited.The situation has been further exacerbated by excess inventory. Even when all inventory is consumed or written off, the exacting demands from the system vendors coupled with fierce competition will ensure that mux/demux prices continue to decline. This is a tough market, with lots of competition, oversupply (such as excess TFF capacity), and four technologies all vying for market share.Technology ProspectsSo, is the condition of the mux/demux market terminal? Well, not quite, but the long-haul market is definitely providing slim pickings at present. Interleavers are a current bright spot, but it is not clear how long this may last.Thin-film filters are attractive to 200GHz metro applications, while bulk diffraction gratings are challenging AWGs on cost (and could become cheap enough to take on the series components) but facing difficulties overcoming market bias. AWGs have overcome this bias themselves and are also attractive for emerging node applications such as optical performance monitoring, gain equalization, and ROADMs. Fiber Bragg gratings and TFFs are attractive for OADMs due to their robust channel spacing and modularity.The Challenges AheadMore than 60 companies are jockeying for position in this market space, pursuing four different technologies. Yet it is expected to be a relatively low-growth market until 2005. Terminal mux/demux is certainly becoming a commodity, but there is no clear advantage to any of the technologies in the growth area of node applications. It seems that the mux/demux players will be severely taxed to navigate through the optical component downturn.