Apr 15, 2002
America's Network

The optical component market will not see a watershed event, but rather continued, steady growth

By Greg Blonder

What will be the next watershed event for optical networking? Everyone with any perspective in communications can remember Ciena and wave division multiplexing. It was an enormously important event, one that established optics as a foundational technology of the Internet Age.

But now? As a venture capital investor in both components and systems and as someone who considers himself an established futurist, I expect nothing dramatic in optics — nothing at all. Instead, I foresee a gradual spread of a second wave of optical innovation whose reach will be subtle and incremental. So beneath our general consciousness will this wave roll on, in fact, that we’re going to look back a few years from now and wonder when all the new optics emerged.

When did it happen?

Think about transistors. When first generation transistor calculators and radios appeared, you actually advertised the number of transistors inside. They were dear things. Similarly, optical components are considered dear today and, in many instances, we still try to minimize their use with electronics.

But that mindset is also changing, partly because a whole generation of engineers has grown up comfortable with optics.

Xebeo Communications, one of my portfolio companies, for example, is building a metro Ethernet switch and not even thinking twice about connecting the planes together with massively parallel optics. There’s nothing novel about it to the engineers; it’s just obvious. Yet three years ago we would have had held absolutely major discussion with teams of engineers forming a red team for electronics and a blue team for optics and battling out a decision.

The “priesthood” of optics is giving way to a believing engineering laity that is ready to look at, for example, the less expensive MEMs switches emerging into the marketplace and say “MEMs switches aren’t so bad, and they will save me a truck roll to reconfigure some cables in a manhole.” Optics has won.

Given this attitude shift, what major changes in optical components can we expect that will infiltrate our networks over the next two to three years?

Much talked about for years, frequency agility will finally happen. Driving the transition are improvements in what turned out to be surprisingly difficult technologies of frequency tunable lasers, greatly increased metro traffic volume from the gigabit Ethernet and the rest of datacomm, and the need for feeds to finally catch up with optical speeds. Morgenthaler has relatively established portfolio optics companies in this space — Agility Communications and FiberSpace — that are getting major design wins even if actual shipment volumes during the telecom capex slowdown are still modest. A third, Inplane Photonics, is just getting started.

The arrival of frequency agility in the metro will give carriers enormous flexibility in provisioning and providing the quality of service that customers require. Carriers will, in essence, regain the ability, lost converting to glass fibers, to watch and control each channel, and finally provide wireline quality in the optics domain.

In addition, by 2004, people will become comfortable performing dispersion compensation and applying electronic error correction code — long the norm in the electronic world — to optics. Optics engineers at carriers have resisted this kind of processing as a contamination of the channel — a case of the optical priesthood, perhaps, staying too holy. But now that the proof is in that such processing is safely doable and offers great cost savings, the tide has turned. In talking with the carriers, I find that they are now comfortable with the idea.

Interest has reached such a pitch that, in one recent week we saw three different business plans for electronic error correction. Once some of these startups receive funding, it will take another 18 months for them to complete trials, and we’ll also need to set some industry standards. But my take on the situation is that we’re seeing activity on all the right fronts. Optical error correction will happen.

When it does, it will allow carriers to use cheaper devices and still offer the same level of quality, while tolerating the outage and failure gremlins that inhabit the real world. If I were to put bounds on the benefits, I would say it’s probably worth about 6 dB in system margin and the ability to hold on to existing fiber for about one or two more generations. And that’s not bad.

Getting together

True integration of optical functions at the substrate level is probably another generation away. Most integration taking place over the next two to three years will be at the systems level — power monitoring combined with error correction coding, for example. (As long as you’re watching bits fly by, you might as well measure the quality of those bits.)

Another kind of integration will take place at the assembly level of optical componentry.

Two companies in our portfolio, NanoOpto and Inplane, are using lithography to create different passive components that lend themselves to automated monolithic assembly. Both will eliminate enormous amounts of hand assembly, with all its inherent costs and inefficiencies. As an industry, we’ve been chasing lower and lower cost assembly countries until we’ve run out of countries. Now, for many crucial components, we won’t have to.

Still another components company in our portfolio, Ignis Optics, is employing assembly methods for transceivers that would have been radical to the optics industry five years ago, but which are now proven (a little too proven, which is why I can't say more.) The result for Ignis and its customers will be much lower cost transceivers for the metro

The cumulative effect of all this non-dramatic optical engineering will be a steady descent down the cost curve. That curve, in my opinion, will never be as steep or predictable as Moore’s Law has been for microprocessors. Optics still depends too much on fundamental breakthroughs in physics. But the foreseeable decline in costs will remain more than meaningful to systems builders and carriers.

According to plan

So, steady–as-she-goes optical innovation and investing from here on out? Not completely. First, the data network continues to grow at 50% per year and seems certain to keep on absorbing new technology. It’s hard to get much more exciting than that. In the current environment, venture capitalists still can’t chance overpaying for a project going in because we won’t be able to oversell it on the way out. But we still find it financially attractive to make new optical investments and to support our existing portfolio companies.

Second, we see strong signs of high drama elsewhere in communications that will add to our interest in optical investing. In my view, for example, the next big communications revolution will come with the rise of public wireless Ethernet. Based on 802.11 wireless and known popularly as Wi-Fi, this form of public wireless Ethernet is set to roar past 3G cellular and take the US by storm.

Just as everyone expects his or her cellular phone to work in every public building, everyone will expect to have Internet access wherever they are. Pockets of access — many ad hoc — have emerged around the San Francisco Bay area, in London and Harvard Yard and are proving wildly popular. It shouldn’t be too long before some established carrier recognizes Wi-Fi’s potential market value and seriously turns it into a major revenue producer. And much of that wireless capacity will be backhauled by an optical web infrastructure- by then it will be the “natural” thing to do.

Many in the industry will find this scenario debatable (and requiring at least another column), but the larger point is this: The spread of optics throughout the entire communications system with all of optics' advantages of speed, bandwidth and ever lower costs is now unstoppable. It thus opens the way to more dramatic innovations from other corners of the communications industry.

Those innovations could include WiFi or they could include widespread voice-over-data transmission, or they could include a dozen other major life-and-business-changing technologies. These will represent new investment opportunities in themselves, but they will also feed back into optics growth itself and create a virtuous cycle. So, despite current industry gloom, we continue to be encouraged about optical component and systems investment.

Greg Blonder is a general partner at Morgenthaler and is based in Princeton, N.J.