x
Optical/IP

ECOC: Back to the Lab?

COPENHAGEN -- ECOC 2002 -- The mood at this year's European Conference on Optical Communications (ECOC) would have been a lot brighter if the focus had been on academic research, rather than the woes of the industrial players, say delegates.

Over the past few years, the rush to commercialize new optical products has meant that a lot of fundamental science issues have been left by the wayside. Now, with the industry in its "quiet period," it's the perfect opportunity to redress the balance, says one university professor with a desire to stay out of the limelight.

Consider the invention of the laser and erbium-doped fiber amplifier (EDFA), he says. Both were the fruits of fundamental research by boffins who had no inkling of how their inventions would underpin the development of a new multibillion-dollar industry. Similar initiatives are needed, he feels, to secure the future of optical technology in five to ten years time.

It must be said that academics can afford to be more far-seeing, being insulated to some extent from the financial boom and bust that's befallen the telecom industry. University and research institute funding is typically awarded in three- to six-year tranches, sometimes longer -- enough to sit out an entire boom-bust cycle.

ECOC, of course, is the place to see both applied and fundamental research in action. Here's Light Reading's pick of the hottest developments presented at the conference.

Hero Experiments 40 Gbit/s was the sweet spot for hero experiments this year.

  • Mintera Corp. claimed to better its own world record, set at the Optical Fiber Conference (OFC) earlier this year (see Mintera Sets Transport Record Again). It says it managed to send 40 channels at 40 Gbit/s (the same as before) over 5,200 kilometers of a new-fangled fiber called UltraWave -- an improvement in the unrepeatered transmission distance of 1,600 km. The key difference between the two experiments seems to be the use of extended FEC (forward error correction) in the latest one -- a technique that adds coding overhead to allow transmission errors to be removed at the receiver.

  • Alcatel SA (NYSE: ALA; Paris: CGEP:PA) claimed a new world record for distance multiplied by capacity. It sent 6.4 Tbit/s, composed of 159 channels at 42.7 Gbit/s each, over 2,100 km. It deployed a technique called "phase shaped binary transmission," which shapes pulses in such a way that they are extraordinarily resistant to dispersion.

  • Engineers from Fujitsu Laboratories Ltd. tried to get a little more in touch with reality by sending 40-Gbit/s signals over standard singlemode fiber (SMF), interspersed with polarization maintaining fiber (PMF) to simulate a link with very high dispersion. Using a widget developed in-house that automatically compensated for Chromatic Dispersion and Polarization Mode Dispersion (PMD), they managed to send 44 channels over six 100km spans .

    All-Optical Wavelength Conversion

    No fewer than six postdeadline papers outlined improvements in devices for wavelength conversion. This represents progress towards making the all-optical equivalent of the electronic repeater. Here are some examples:

  • Working with JDS Uniphase Corp. (Nasdaq: JDSU; Toronto: JDU), Eindhoven University of Technology-COBRA has developed a semiconductor optical amplifier (SOA)-based wavelength converter monolithically integrated with a digitally tunable laser, which can convert the input signal to any of the wavelengths available from the laser.

  • Scientists from the Optoelectronics Research Centre in Southampton described a tuneable wavelength converter based on cross-phase modulation in holey fiber with enhanced nonlinear properties.

  • Engineers at Corning Inc.'s (NYSE: GLW) research center in the U.K. (formerly BT Labs) described an SOA-based wavelength converter that can be fabricated using passive assembly techniques. They also presented data showing regeneration of a 40-Gbit/s signal using the device.

    Other Components

    A number of papers described advanced subsystems such as dispersion compensators, reconfigurable add/drops, and wavelength selective crossconnects. In addition, there were advances in making long-wavelength Vertical Cavity Surface Emitting Lasers (VCSELs), a type of laser that is inherently cheaper to manufacture than standard DFB types.

  • Swiss startup BeamExpress Inc. published its first result from its 1550nm tunable VCSELs.

  • VertiLas GmbH reported the first data transmission results from its 1550nm fixed-wavelength VCSELs.

    — Pauline Rigby, Senior Editor, Light Reading
    www.lightreading.com
  • lightbridge 12/4/2012 | 9:46:33 PM
    re: ECOC: Back to the Lab? Why is a startup like Mintera spending huge money (forty 40Gb/s transponders) towards the world record race? Shouldn't they hibernate until 40Gb/s picks up? It appears, they still have the Bell Labs mentality of their founders, which might be good for the scientific community, but is a problem when you have limited resources.

    lightbridge
    light4me 12/4/2012 | 9:46:32 PM
    re: ECOC: Back to the Lab? I am a bit tired of reading that VCSELs are cheaper than DFBs. Since most of the cost of the modules on the market come from the box (usually gold-plated), the TEC, the optical isolator,... I challenge anybody to demonstrate that what the customer buys (a packaged laser) is cheaper in the VCSEL flavor than in the DFB flavor (of course we are talking ~1.55 microns).
    eer2000 12/4/2012 | 9:46:28 PM
    re: ECOC: Back to the Lab? Without having done a detailed cost estimate, here are a couple areas where VCSELs could be cheaper than DFBs:

    1) bar sort: VCSELSs could be fully characterized on wafer; DFBs must be singulated into bars, stacked and loaded into a fixture, and coated, unstacked, and tested a bar at a time

    2) optical train: VCSELs have a larger, more circular optical mode than DFBs, which could significantly reduce optical train packaging costs, especially if a reliable passive alignment scheme is used (which may not be possible for DFBs due to small mode size)
    LightBeating 12/4/2012 | 9:46:27 PM
    re: ECOC: Back to the Lab? eer,

    Agreed. The chip itself of course costs nothing, VCSEL or otherwise. Testing and packaging are where the costs are (apart from the other materials themselves), and VCSELS might end up having a slight advantage there. On the other hand, they lack the power. Still, I find the progress on VCSELs more exciting than the other stuff talked about in the article: more distance*bandwidth records, stuff like SOA or fiber based wavelength converters: old concepts that never really worked. Cheap tunable lasers will give you the wavelength flexibility you need if you ever need it, wavelength converters will never make it as a technology.

    Personally, I don't see much exciting "fundamental" science these days. Nothing like the EDFA, and even that was recycling of very old stuff done in the early 60's. Fiber optics is a mature field. There's still plenty of room for clever engineering, but new science? Nah!...

    LB
    st0 12/4/2012 | 9:46:23 PM
    re: ECOC: Back to the Lab? (1) not all the dfb are made equal. some can be tested on wafer as well.
    (2) with recent lens development, the beam profile can be corrected without too much cost.

    Give a year or so, DFB is the way to go..my 2 cents.
    ----------
    Regarding the University research, it is sad that corp long term research is almost extinquished.

    The deficiency of the university research is that poor development and understanding of fabrication process. The university established so call research "product", usually need substaintial development and many interation in order to become product (sometimes, complete re-vemp). Either due to lack of understanding by the industry about the fundamental of the research, or due to lack of practical experience of the university academics. Being said that, some of the academics who defected from the corp. research were the best.

    On the other hand, some of the university IP office is really scary. You got have strong heart to sign those 10 point font 4 - 6 pages NDA... worse than licencing the patent.. particularly, you not even know if the stuff is going to do the trick you want it to... (some of patent application just "idea"...let's not go there). All of them want you get long term commitment, even down payment, before even talk to you...

    --st

    Unfocused 12/4/2012 | 9:46:21 PM
    re: ECOC: Back to the Lab?
    Why are these Minteras guys spending on 40 Gb/s instead of hibernating ...

    ---

    I spoke to 3 of the VPs at OFC 2002. They seemed to be aware of avoiding to spend too much like OptiMight supposedly did on their MCI fieldtrial. Without having seen the ECOC article, I guess they are just squeezing the same loop as they had at OFC (FC7) + adding the FEC which is not too costly when you hibernate anyway.
    LegoAss 12/4/2012 | 9:46:16 PM
    re: ECOC: Back to the Lab? Industry pays for tradeshows and not academia. If the show shifts towards academic research, industry (and their money) will go elsewhere like NFOEC and/or Supercomm.
    HOME
    Sign In
    SEARCH
    CLOSE
    MORE
    CLOSE