With the deployment of 40-Gbit/sec systems pushed back 18-24 months, optical test equipment vendors focus on improving existing devices for near-term applications.
By MEGHAN FULLER
In times of economic prosperity, new car sales generally soar. But when the economy takes a turn for the worse, people often put off buying their dream car, opting instead to get more out of the cars they already own. However, that may mean getting the car serviced more often-checking the engine, transmission, belts, and brakes. Routine testing is more important than ever.
The good news for optical test and measurement equipment vendors in the current economic downturn is that there is always a base level of business. Carriers and service providers may have curbed spending on new infrastructure, but they must maintain their existing infrastructure, and that means continuous testing. Similarly, component and system manufacturers must evaluate new designs and keep production outputs steady.
That said, optical test equipment vendors recognize that in a perfect world, their customers would rather do without them. "One of the things I tell my engineers once they are hired and start working for me is that their customers will never love them," admits Stefan Löffler, product marketing manager of passive-component testing at Agilent Technologies' Optical Communication Measurement Div. (OCMD-Böblingen, Germany). "And that is simply because the other paradigm is that testing, by definition, is a pain."
Easing this pain has become the test equipment vendor's biggest challenge in the current economy. Their ability to deliver easy-to-use, cost-effective systems to enable their customers to lower costs and improve manufacturing yields has and will continue to help them weather the economic storm-though the test segment seems to be faring better than others. According to market researcher Frost & Sullivan (Mountain View, CA), the worldwide fiber-optic test equipment market will jump from $1.7 billion in 2001 to more than $6 billion by 2007 (see Figure).
While the deployment of new test equipment for 40-Gbit/sec transmission systems will be a key driver of that growth, this market is not expected to materialize for 18-24 months. In the near term, test equipment vendors are hard at work to improve devices for existing applications.
Mainstream fiber installation and maintenance test equipment is not becoming obsolete; technicians still need devices like power meters, optical time-domain reflectometers (OTDRs), and optical spectrum analyzers (OSAs) to characterize dark fiber and, in some cases, existing fiber. Thus, test equipment vendors have focused on improving these devices. End users are asking for equipment that is easier to use and delivers faster cycle times, increased throughput, and automatic data analysis. In short, they are looking for "a brick with a button," admits Peter Schweiger, channel manager of Agilent Technologies' Optical Network Test (ONT) division (Böblingen, Germany).
One-button devices are necessary, agrees Jerry Gentile, business unit manager for Acterna's Transport Div. (Germantown, MD), because "you always have lower-skilled technicians and a lot of turnover in that large work force, so we've developed instruments on the portable side that are very easy to use." Among these instruments is the TestPad 2000, which boasts an easy-to-use touchscreen, automated testing features, and customizable scripting capabilities.
Another way to increase efficiency is to combine several devices into a single test platform capable of measuring multiple parameters. Stéphane Chabot, business unit manager for outside plant products at EXFO Engineering Inc. (Quebec), recalls speaking with several customers who were using one test device to do optical loss testing, another device to do OTDR testing, and another device-such as an OSA or multiwavelength meter-to do DWDM testing. They also used separate devices for polarization-mode dispersion (PMD) and chromatic dispersion (CD) and still others for protocol testing and bit-error-rate (BER) testing.
EXFO responded to this need with the field-based FTB 400 Universal Test System (UTS), launched last June.
The advent of DWDM in the mid-to-late 1990s continues to affect the test equipment market. "The field test area went from being OTDR-driven to OTDRs plus many other pieces of test equipment needed for turning up DWDM systems," explains John Chapman, president of the Optical Div. at NetTest (Utica, NY). "And that means dispersion testing, for one."
PMD and chromatic dispersion cause light signals to become unreadable over long distances or at high data rates. At rates up to 2.5 Gbits/sec, dispersion is not really a problem, contends P.J. Kleffner, business development manager at Tektronix Inc. (Beaverton, OR). But at 10 and 40 Gbits/sec, dispersion testing is an absolute necessity.
"Test tools that are being used more now because of 10 Gbits and 40 Gbits are things like optical spectrum analyzers that can test the device's responsiveness to PMD, to be able to characterize for it. That's a major tool," says Kleffner.
While such testing is critical for long-distance transmission systems, even in the short-haul it is often necessary. Chris Brozenick, director of product marketing for Acterna's Transport Div., recently spoke with a customer installing a metro network. "They want to put DWDM in, but what they are finding, unfortunately, is that the majority of the fiber that they have in their footprint will not support more than four to eight channels," he explains. "They really would like to support 16-32 channels, so it's forcing them to do a lot of testing of the fiber in the metro to see what it can actually support. They use things like OTDRs, OSAs, and other optical test sets to check that out."
With the long-haul segment significantly built-out, carriers and service providers have turned their attention to metro applications. According to a recent report from Pioneer Consulting LLC (Boston, MA), the worldwide metro optical-network market is expected to jump from $2.9 billion this year to $13 billion by 2005, which presents new opportunities-and challenges-for test and measurement vendors.
Unlike the long-haul, the metro network encompasses multiple layers, protocols, and standards, which requires a shift from "homogenous BERT-type testing to more complex testing," asserts Acterna's Gentile.
"In the long haul, SONET ruled everywhere," explains EXFO's Chabot. "In the metro, we still see SONET and SDH everywhere, but in the past couple of years, we've seen a lot of IP-based networks, ATM, IP-over-ATM, etc. And those mesh networks are competing with the actual SONET ring-based architecture. These networks have different topologies, which present different challenges in testing."
In long-haul networks, links are typically 150 km, with an amplifier or repeater every 80 km or so. Metro networks, in contrast, are ring-based and may include a fiber-optic "event," as they are called, at random points throughout the link. "Events can be very close to one another," explains Chabot. "You can have four events 20 m apart at kilometer 16, and then another three at kilometer 32, and so on and so forth. You still need the same dynamic range, but you have to detect events that are a lot closer to one another."
While the metro requires basically the same type of test equipment, that equipment must now be tailored to suit the metro environment. If technicians use a standard OTDR for this type of application, they would see one merged event instead of several different events.
The solution, contends Chabot, is to improve the OTDR's specifications. Previously used to measure loss and locate breaks or misalignments, today's OTDRs test additional parameters such as CD, which is a critical problem at OC-192 transmission speeds as well as in previously installed fiber that may not have been checked for dispersion. Dynamic ranges have increased, and the speed of test has been improved. In the past, OTDRs took between 30 sec and 3 minutes to characterize fiber; today's devices take just 10-15 sec for the entire acquisition.
The test and measurement market from the optical-component side has been harder hit by the current economic downturn. According to Mark Fishburn, vice president of technical strategy at Spirent Communications (Rockville, MD), the current economy has changed the business strategies of many companies focused on component testing. Component test equipment vendors now must differentiate between what their offerings can and should test from a technical standpoint and what they can and should test from an economic standpoint. Often, they are not the same.
"If you think about testing a three-port switch on all 80 or even 10 lambdas at the same time, you would probably need 30 pieces of OC-192 test equipment to analyze each and every packet," explains Fishburn. "It might cost $20 million to put such a testbed together. This is clearly not a viable thing to do in the marketplace right now."
Moreover, Fishburn contends that no such equipment even exists to test 80 lambdas at 10 Gbits/sec to ensure that every packet arrives in the right place at the right time and to do BER testing and physical or optical testing across those networks. And that, he says, is the challenge facing test equipment vendors in the component space-"to decide how much further they should go ahead and create the very complex test systems needed to make a thorough test of these networks."
The end goal for anyone in the telecommunications space right now is to reduce the cost per bit. To that end, Agilent's Löffler believes test equipment vendors must be committed to fostering technological improvements and bringing them to market-but in a cost-effective manner. "There is a great belief that with every step, every leap forward in the transmission capability and capacity, the cost per transmitted bit will be significantly reduced," he says.
When asked to predict the next such leap forward, Löffler does not hesitate to name 40-Gbit/sec transmission systems. Though such systems will require more sophisticated test equipment, component manufacturers do not want to pay more for this equipment, which, according to Löffler, is another big challenge facing today's test equipment vendors.
Because 40-Gbit/sec systems are so susceptible to dispersion, each component must be tested for dispersion as well. Take, for example, a WDM filter, says Löffler. A few years ago, that filter may have cost as much as $200, with a test cost contribution of about $20. Today, component vendors would like to sell that filter for $20, one-10th the previous price. "The test cost contribution of this piece should be far below that and should be a smaller percentage than before," contends Löffler. "Let's say about 50 cents. Now, how can we achieve that?" The answer, he says, is to improve throughput or the number of units a device can test per minute per station.
A component manufacturer expects to achieve a certain throughput each year, and from this projection, test equipment vendors can predict how many test devices that manufacturer will likely purchase. The amount of test equipment these manufacturers purchase must be calculated as part of their return calculation and therefore affects the price they ask for the component. Test costs can be as much as 20% or 30%, says Löffler, so if test equipment vendors can reduce the cost of equipment as well as the number of test devices required, they can improve the component manufacturers' cost structure.
"So if you increase the throughput, you automatically put the same factor, roughly, into the test contribution," he explains, "which means 10-fold throughput is a 10th of the cost contribution. Similarly, if we measure three test parameters with one test set rather than three test parameters with three test sets, that's a threefold improvement of the test costs."
While it's too early to predict the next DWDM-like improvement, tunable lasers may receive a great deal of attention in the coming months. The most expensive part of a test system is the tunable laser, which has led several test equipment vendors to devise a solution whereby each laser may be shared among several workstations. Until now, each workstation must conduct exactly the same test. However, the near future-perhaps even as early as this year-may bring the advent of asynchronous test measurements in which one laser may be shared among several workstations, with each station capable of conducting a different test.
"This means that if you have four test stations that use the same laser, you can split the laser among four test systems and reduce the price contribution of this laser by a factor of four," contends Löffler. "This is simply an investment you have to make as a manufacturer, but it's one big part that contributes to cost reduction."
Adds NetTest's Chapman, "As the market gets tight and things get more difficult, prices come under pressure."
Before the recent economic downturn, test equipment makers focused on reaching the next technology plateau. They were proactive, but now the market is perhaps best characterized as reactive-to both the economy and its customers' needs.