Article Info

Future of Optical Networking – Lightreading Conference

I attended the Lightreading “Future of Optical Networking” conference in New York last week.

I had high expectations for the conference based on one I attended two years ago on Ethernet in the WAN. At that conference there were a number of participants from start ups and established companies who led a vigorous debate about what the future of SONET/SDH looked like. Even though I lived and breathed Ethernet over SONET/SDH at the time, and met regularly with the companies on the panel, the debate that ensued was highly educational and enlightening.

The recent conference was oriented more as a PR event. It was clear that Cisco, Fujitsu, and Ciena paid Lightreading to assemble an audience (and at one point a speaker joked to this affect) of carriers and financial types so that they could preach their gospel. There was some good point-counterpoint on ROADMs (Reconfigurable Optical Add/Drop Multiplexers) but for the most part the panel sessions were more lecture and less debate. I think this conference was an anomaly, as I feel Lightreading research and reporting is superior to the work done by ‘more reputable’ firms like iSuppli and Dell’Oro.

I also had the good luck to win the Apple iPod Nano given away at the conference conclusion. I found this particularly ironic given my vocal opinions on Apple and iTunes. Another member of my family harbored no such opinions and quickly nationalized the asset faster than a South American Dictator($$$ WSJ Link). Thank you very much Lightreading.

Joe Weinman, VP of Strategy and Emerging Services at AT&T spoke for nearly an hour on future applications for the net. He positioned AT&T as delivering a future where they operate a myriad amount of services on the net – television, compute farms, remote storage, even RFID tag reading. AT&T doesn’t strike me as having 1/10th of the corporate agility to implement all of these programs. How will AT&T outpace other smaller companies that offered the same services over AT&T’s network?

Regardless, Joe was a very good speaker. If AT&T has execution equivalent to this guys vision, they’ll do all right. I engaged him afterwards a question on Project Lightspeed- “What is the number one reason a cable customer would switch to AT&T for Video Services” – and didn’t get a cohesive answer. AT&T still needs to communicate this clearly to the market.

Scott Clavenna presented some Heavyreading forecasts for the optical equipment market.

Market Size 2005 -> 2008
Metro WDM $1.25BB -> $2.2BB
Long Haul Transport $1.4BB -> $1.65BB
“Legacy” Multi-service SONET/SDH $3.5BB -> $4.2BB


  • Metro WDM is going to grow like gangbusters driven by Carriers deployment of video services. Most of the growth will come from Ethernet transport. The equipment in this sector tends to be the more ‘unconventional’ equipment like OTN transport or Ethernet over wavelength.
  • Long Haul Transport grows slower, but is driven by broadband growth at the edge. Capacity constraints are starting to show up.
  • Legacy SONET/SDH stops growing in 2007. Revenue is driven by wireless backhaul of 2G/3G services and upgrading existing MSPPs to support Ethernet linecards.

I think market research firms should refrain from forecasting hard numbers, and that these efforts should be left to those who have skin in the game. Here’s my take.

  • The impact of carrier video on WDM transport is overstated. Optical transport will grow over the next five years, but driven by consumer broadband usage growth and enterprise services growth. It isn’t video distribution – most video will be cached at the very edge of the network to reduce transport costs and more importantly- preserve customer QoE (Quality of Experience). It will look like broadband data. That’s why we’ve been investors in Akamai (AKAM) and continue to own the company at these high valuations.
  • Long Haul Capacity constraints are indeed showing up, particularly for people seeking large 2.5Gb/s connections through the core. After the bubble, no one has spare capacity in their network if they can avoid it.
  • Legacy SONET/SDH may stop growing, but an examination of the spending will show dramatic changes. Spending will shift from TDM grooming and copper DS-3/DS-1 connections towards Ethernet over SONET/SDH and L2 switching, and from OC-3/12 tribs to tunable OC-48 interfaces. The only thing ‘legacy’ about the spending in this category will be the fact that the old sheet metal boxes are hosting these new advanced service linecards.
  • In North America, the growth of packet over wavelength services will be muted because the majority of capex continues to be from Carriers with major legacy SONET/SDH networks. While Heavyreading is forecasting big growth in advanced WDM systems, they also point out 57% of carriers will use existing equipment. That’s 57% of 60 polled carriers, not 57% of capex. You can assume that 20 of those 60 carriers are 80% of the capex, and they have a lot of legacy hardware.
  • The real growth for packet over wavelength will be in Asia as well as hero projects like BT’s 21CN. The most uncertain aspect of metro WDM growth isn’t just how big it will be but who will emerge as the market leader. We still believe that Huawei will dominate next gen WDM much like they have captured market share in DSL, and that Ciena (CIEN) investors may not have factored in this outcome.

Lightreading’s account of the conference is summarized here.


Comments are disallowed for this post.

  1. Hi,
    I read your comments on the Metro market, and agree with you that video is not driving the expansion right now, in fact storage extension for enterprise networks as well as government spending is driving a lot of the metro demand, most people just don’t see it. I’ve got some data on the U.S. enterprise verticals deploying optical networks, and some ‘hard’ forecast numbers as well if you want to contact me.
    Eve Griliches

    Posted by Eve Griliches | April 27, 2006, 4:29 PM
  2. Please do share…

    Posted by Andrew Schmitt | April 27, 2006, 6:17 PM
  3. Hey Andrew,

    Thanks for the write-up. good stuff and glad you could make it.

    On the forecasts, we actually started Heavy Reading with a mission NOT to do forecasts – we figured numbers were rather discredited after the bubble, as you point out, and it was time for more qualitative research. I think we were right, and we got some great analysts here to go right to the source and do surveys of service providers, customers, and then talk to vendors and get a sense if they were on target or not.

    I ended up relenting on optical because as it has started to truly grow again, I find numbers are an important way to tell that story, at least part of the story. I put together the numbers over the past four months by getting reports from all the vendors about their CY 2005 revenues, including great detail from Huawei, then went out and met face to face with major telcos and MSOs. We’re still going to put out qualitative analysis, but the numbers help flesh out the whole story.

    I agree enterprise and govn’t are great drivers for metro DWDM today, and provide much better margins than telcos, but when we forecast out to 2008, every major operator with massive capex told us video was driving their optical networks growth, particularly for regional networks. I didn’t find any real coherence on just what these optical architectures would look like, though obviously Ethernet/packet over WDM is a common theme. But the big ROADM RFPs out there in the telco and cable MSO space are explicit about video carriage.

    I’m with you on Akamai, caching has to be a big part of this, but for now I’m sticking with video as a driver for 2007-2008 in the WDM market because the plans are in place. that said, it’s quite possible that “video” is just a sexier way of saying “broadband” now, so we may really be talking about the same thing.

    On the Sonet forecast, I went back and checked the survey out and filtered the results to just show those operators with capex of $1 billion or more annual. The pie chart was about the same – evenly split between expanding with current gen of multiservice Sonet, deploying new generations of MS Sonet, and capping Sonet in favor of packet over optical. So the capex weighting doesn’t yield a big change in the numbers, though I will warn that these surveys are not always taken by the person holding the pen over the check. I’m totally in agreement with you about how the spending on MS Sonet will shift away from TDM to OC-48 and Ethernet. No doubt abougt that at all.

    In placing bets on who will dominate LH DWDM, you have to go with Infinera and Huawei. They are both differentiated in their own way, and quite disruptive. Both are game changers like none other.


    Scott Clavenna

    Posted by Scott Clavenna | April 28, 2006, 9:27 AM
  4. Scott:

    Thanks for the detailed response and follow up. I’ve always recommended HeavyReading to others exactly because there is no forecast.

    Going forward, if you do forecast, it would be good to adopt a format where you track accuracy. This would help people see trends in your projects and add value. It also shows that you’re willing to be accountable, something most market research firms avoid like the plague.

    Hope to see you soon.


    Posted by Andrew Schmitt | May 4, 2006, 2:19 PM
  5. Trackbacks / Pingbacks

  6. WDMBlog » Lightreading Future of Optical Networking Conference Wrapup | April 26, 2006, 12:14 AM
  7. Infinera Blog » Scott Clavenna Calls It | May 10, 2006, 5:16 PM