Optical components

Report Details Component Hot Spots

Components that help carriers make the most of installed fiber will succeed in the coming months, while those aimed simply at boosting raw capacity will collect dust for at least another two years.

So says a report titled "Component Conundrum," published this month by Light Reading's subscription service, Optical Oracle.

The report posits that carriers want equipment that lets them carry more data on installed fiber. Because of this, arrayed waveguide gratings (AWGs), interleavers, and some types of tunable components are likely to sell relatively well over the next 12 to 24 months. In contrast, 40-Gbit/s transport chips are "at least two years away" from market penetration of any significance.

"What carriers want and need right now are lower costs and improved efficiencies," says Chris Bulkey, Optical Oracle research analyst and author of the report. That means more channels carved out of existing fiber, not faster pipes.

AWGs are key to creating these channels. AWGs have succeeded in dominating today's 40-channel DWDM market because, according to the report, they offer lower losses than thin-film filters for this application. To reach 80 channels and above, however, AWGs will need a boost from interleavers. Companies that make these components include Avanex Corp. (Nasdaq: AVNX), JDS Uniphase Inc. (Nasdaq: JDSU; Toronto: JDU), and Oplink Communications Inc. (Nasdaq: OPLK).

Tunable components also are hot, because they promise a way to make the growing number of DWDM channels more manageable. Specifically, tuning can be used to add, drop, or switch channels via software. And some forms of tunable components open the possibility of routing by wavelength -- a dramatic improvement over today's routing methods, which rely on electronic mechanisms.

But tunable components will likely grow more slowly than AWGs in the near term. According to the report, carriers need to be shown that systems based on tunable lasers and filters can be deployed cost effectively and meet performance expectations.

A number of manufacturers, including Agere Systems (NYSE: AGR), Bandwidth9 Inc., Fujitsu Ltd. (KLS: FUJI.KL), Marconi Communications PLC (Nasdaq/London: MONI), and Nortel Networks Corp. (NYSE/Toronto: NT), are working to convince carriers (via system OEMs) that tunable gear is viable. What's more, integrated optical add/drop multiplexers equipped with tunable filtering capabilities are starting to emerge in products from the likes of Bandwidth9, Cidra Corp. (Nasdaq: CIDC), and Solus Micro Technologies Inc..

While these components are likely to flourish, others, especially 40-Gbit/s transport chips, will be delayed. Carriers are seeking to mine the capacity they've already built out, and the falling cost of 10-Gbit/s systems is convincing them to focus on that data rate as the backbone standard. According to the report, it will be at least two to three years before carriers seek to up the backbone ante with new baseline speeds.

— Mary Jander, Senior Editor, Light Reading

Editor's Note: Light Reading is not affiliated with Oracle Corporation.
<<   <   Page 2 / 2
realoptics 12/4/2012 | 7:53:00 PM
re: Report Details Component Hot Spots To: Milano:

FYI, nobody is joking on athermal grating and it is not from a market report either, it is from a simple fact. By the way, when we saying grating, it is not AWG.

BaySpec, inc. in Fremont, California is deliverying athermal, volume phase grating based 40 channel mux/demux to its customers, the devices are athermal over 0-70 degrees C , unlike the pseudo ones you mentioned on NEL's AWG. I know again for a fact that the company will be hpappy to send you a 40 channel sample if you make a request to BaySpec, Inc as a system integrator.

When Nortel is using Avanex interleaver, it does not necessaryly means their interleaver is better, it may just mean the customer may not have known that there are better things already out there to defeat the whole purpose. When we were visiting NFOEC, we happen to see BaySpec's booth besides Avanex's, huge, but empty(in terms of products) booth, both BaySpec and the other were showing a 40 channel mux/demux, your favorite interleaver guy was showing a 6.5 dB insertion loss with a 2.0 dB uniformity, while BaySpec has a device with a 2.5 dB insertion loss with a 0.5 dB uniformity(Avanex engineers were shocked and shamed, as those good Chinese guys would feel after they themself found out the direct comparison), the results convinced people that you would only need one box to achieve even better performance, so why would you still want to go for 3 boxes to have even much worse results? So interleaver or not?

The Chinese says that 'one does not need to take out the pants in order to fart'. Interleaver is doing the same type of unnecessary things to the mux/demux technology, the grating technology can do the ultimate high channel count directly if one knows how to do it, and it is already hiting teh market. It is athermal, scalable, compact,and really provides low cost per channel.

Petabit 12/4/2012 | 7:52:50 PM
re: Report Details Component Hot Spots Good discussion, poor article, but you are missing some key points.

Most systems today are still deployed as point-to-point routes. You need one mux at one end and one demux at the other. Adding OADMs or PXCs doesn't really change your need for muxes. At some point this system will have one channel on it, and at a later time it will have more channels. The challehnge is to come up with a mux architecture that scales cost effectively from one to ~100 channels.

AWGs have historically lost out because of their high initial cost - all those fibre alignments really pump the cost up. The same is true for the other technologies that give a single mux: echelle, VIPA, etc.

Today the market is dominated by TFF and FBG, the reason comes down to the fact that they are cheap in ones, and not too expensive in many.

Interleavers are an interesting case in point. In some senses they perform a mux by joining together odd and even channel streams. However in order to produce this flexibility they have a much narrower passband than some of the other technologies. They are used in systems where the channels are not too tightly packed, or are bidirectionally interleaved (e.g. Nortel). They are not about to take over the world.


PS: Mr Realoptics, "By the way, optics is something not easy to predict, we consider 5-8 universities in the U.S. have the best optics programs(you know the names) in the world, and some others are in Canada and China(totally less than 4), if the analysts were not taught in one of those institutions, it is hard for them to do a super job."

That's a pretty arrogant position. Especially when you consider that most of the really famous optics gurus were educated in Europe - you company might do a little better to look further afield. Oh, and by the way, all the best optics people I have ever worked with, none of them went to a formal optics programme - they all figured it out for themselves.
realoptics 12/4/2012 | 7:52:35 PM
re: Report Details Component Hot Spots The one sentance I do agree with Petabit is that interleaver is not going to take over the world!

Indeed Optics is an 'arrogant' field. But please note that I did NOT say: "that most of the really famous optics guys were educated in Europe", as you were trying to sugget to the other readers, I was saying the best universities are in the U. S. (5-8) universities.

By the way, not trying to be arrogant again, the best optics people you have ever worked with, "none of them went to a formal optics programme - they all figured it out for themselves", that could not be true!-It may be even an insult to the optics field for which only the amateurs would consider it as an easy subject. Therefore we would tend to think the people you ever worked with are not necessaryly the best optics people.
<<   <   Page 2 / 2
Sign In