It used to be that fiber optic technology, when it first garnered its "state-of-the-art" moniker, was something many cable professionals thought was on par with "Flash Gordon" or "Star Trek" (depending on your age). But today, there aren't many engineers who would consider upgrading or building a broadband network without some amount of glass cable.

While the cost of boosting fiber content in networks will never be cheap, the demand for more capacity is beginning to grow. Such things as high-speed data, near-video-on-demand, video-on-demand, and IP and HFC telephony are spreading out on cable's palette of services.

Yet, when it comes to fiber optics, engineers are loathe to go in and rip up what's already in place. Riding to the apparent rescue are new fiber optic technologies, in both hardware and software. There is also a renewed interest by both vendor and operator alike in dense wavelength division multiplexing (DWDM). With these new developments, fiber optics has taken on a renewed lustre and is once again on the cutting-edge of technological advancement in the broadband communications industry.

Wavelength envy?

Dense wavelength division multiplexing will be TCI's "technology of choice" for new fiber construction down to the secondary hub level, says Oleh Sniezko, director of transmission engineering for TCI Communications Inc. Sniezko says that TCI has been driving the industry's manufacturers to apply DWDM technology to analog transmission, and thus "significantly simplify" the MSO's cable network.

"All the processing equipment will either be in the headend, or at the customer's home," he explains. "That way, the network is simplified, because it is fully transparent to whatever services you want to introduce."

Sniezko says TCI was due to test one operational DWDM system in the latter part of January, and another in the second week of February, at another manufacturer's facility. He expects that by the second quarter of this year, manufacturers will be able to provide a DWDM system that can be deployed in the field, which will also be scalable.

"We are riding on a very steep part of the price curve today because this is a new technology," he says. "We expect that by making (DWDM) scalable, we"ll ride on a shallower part of the curve in two to three years."

One of the industry's most aggressive operators when it comes to state-of-the-art technology and fiber deployment is Time Warner Cable (TWC). And, while the company continues its break-neck pace on HFC rebuilds around the country, Paul Gemme, TWC's vice president for plant engineering, says the company still has some questions about DWDM and where it fits into the overall picture.

"We've been upgrading our plant now for the past several years," says Gemme. With more than 200,000 miles of fiber plant, most of which is 750 MHz HFC to 500-home nodes and fully two-way, the company is close to having 50 percent of its total plant mileage upgraded."

Often, Time Warner's solutions end up being a conglomeration of different vendors' equipment. "We tend to take the best of what's available out there," says Gemme. "We try to package it together in our own fashion. Sometimes that entails using several manufacturers because nobody's got a corner on the brain market and the technology is pretty close.

Osicom's GigaMux DWDM

"But we do testing, and we have an approval process. So, based on who's meeting the specifications and it's usually someone different for 1550 nm (equipment) vs. 1310 nm. It's all over the map. You've got to be continually looking at it and understanding that you will change vendors as new projects come on."

Gemme and his TWC cohorts are keeping their collective eye on technologies like DWDM and asking some basic questions, while trying to keep their options open at the same time. "Everybody is always asking whether we're going to run out of reverse bandwidth or forward bandwidth first," says Gemme. "I don't think anybody knows. So, we want to be prepared for both. To hedge its bets on these types of services, as well as high-speed data, Gemme says the company will continue to run six fibers per node "so that we could ultimately break down node sizes from 500 homes, on average, to probably 150 to 200 homes."

This concern for forward bandwidth capacity, says Gemme, is having an impact on his deliberations on DWDM. "We're looking at it and saying, 'Where does it fit? Does it fit in our architecture? Does it allow us to put (in fewer) fibers per node?' We're not sure yet. We don't think that using 1550 for distribution for the nodes, for example, makes a lot a sense because you're broadcasting — and the ultimate goal is to narrowcast or have more bandwidth per home passed.

"Typically, we're doing that by using 1310 and only splitting our transmitters two to three times. The pricing is coming down on that stuff. So, we're buying transmitters today that we only need to split twice to be as cost-effective as we were three years ago. We are getting to the point where narrowcasting or providing more bandwidth in the forward path per home passed is becoming a reality."

While Gemme hasn't reached a conclusion on DWDM, he's a true industry veteran in noting that when it comes to so-called state-of-the-art technologies and money, he's never one to say never. "It's always the game of economics," says Gemme. "You know, which technology is crossing that economic line first. And, right now, we still think fiber is pretty reasonably priced. We will be testing a lot of WDM stuff this year. It could be that mid-year I'll be telling you a totally different story if things work out the way I'd like them to."

DWDM picture getting crowded

A number of traditional broadband equipment manufacturers showcased new dense wavelength division multiplexing products at December's Western Show, signaling increasing MSO interest in the technology.

ADC Telecommunications rolled out a new 16-channel dense wavelength division multiplexer/demultiplexer, as well as its new OTAU (Optical Test Access Unit). As MSOs create more superheadends, says Jeff Korkowski, director of product management for ADC Telecommunications Inc.'s fiber products group, DWDM is suddenly making a lot of sense. "If you are going short distances, for example, from a fiber optic transmitter out to a node, you would not use DWDM," he notes.

That's because the real cost savings occurs when you can multiplex four or eight channels and thereby avoid having to use repeaters to overcome the signal losses associated with long links. "If an operator can multiplex four or eight channels, and instead of using a repeater, use an amplifier, he can amplify all four or eight units with one unit," he says.

In addition, as cable operators move toward offering new services like telephony, data and PCS, it's more economical to use DWDM, says Korkowski, than to put new fiber in the ground to ease capacity crunches. While ADC is currently shipping the product out for "lab testing," it should be available in quantity sometime this July, says Korkowski.

ADC's version of the DWDM utilizes fused biconic taper technology and Fiber Bragg Gratings to separate, or demultiplex, the signals. The grating acts as an optical filter would, reflecting certain wavelengths of light and allowing others to pass through.

Harmonic's PWRBlazer

"The elegant nature of the design is that the light remains within the glass structure at all times," says Korkowski. "We haven't glued filters onto little devices, or tried to precisely align these pieces."

ADC's OTAU, designed to help operators remotely test fiber systems, is a multiple fiber switch that sits in a headend or remote site and makes the physical connection between any one of the fibers and the test set.

"Fiber counts will go up, and the time is coming when an operator will need the ability to quickly determine where an outage is, and direct a crew to that specific area," says Korkowski. And, he adds, the number of people who understand what is on any particular fiber is minimal. Besides further automating the testing process under normal conditions, an OTAU could save valuable time in the event of a fiber cut, he adds.

Synchronous Group Inc. unveiled its new 1x32 Dense Wave Division Multiplex System at the Western Show as well. Vince Borelli, chairman for Synchronous, believes DWDM is an important development that not only piggybacks on current technology and meets future capacity needs, but increases system value as well. "With DWDM," says Borelli, "you probably would never have to do more than 32 wavelengths. If you did more than 32 today, you're looking at some phenomenal data capabilities.

"A lot of the cable operators put anywhere from four to six fibers, on average, to every node they go to. And they're using 1310 nm. But all of that 1310 has nothing to do with this 1550. You can wave division multiplex 1550 on top of that 1310 all day long. And it won't interfere with anything at 1310.

"But, the point is, if you want to invest in an optical plant and all you want to do is 1310, and that's all you think you can do, then it has a certain (i.e., fixed) value. But, if all of a sudden you can do all this other 1550 data trunking, obviously, the plant just goes up in value."

To help its DWDM effort along, Synchronous also debuted its new Sirius-D digital Erbium Doped Fiber Amplifier, which has been specifically designed and tested for digital applications. It also introduced its new Spectrum Series Digital Transmission system, which provides uncompressed digital transmission. Available as a standard 16-channel per fiber system, the introduction of the new transmission technology allows operators to carry up to seven 16-channel systems on a single fiber.

Scientific-Atlanta Inc. took the wraps off of a second major component in its end-to-end solution for transporting two-way digital and analog services over a single fiber. S-A's Dense Wave Division Multiplexing platform is the newest addition to its Prisma Optical Networks family of optoelectronics products which is currently under development and scheduled for release in the second quarter of 1998.


Despite the technology's bad rap as being too expensive, S-A's Vice President for systems marketing, Lee Johnson, thinks DWDM is a technology whose time and price has come. "We feel like," says Johnson, "from a customer standpoint, it's a cost-effective solution because it allows you to get your revenues sooner. You can deploy a system like this a lot quicker than you can lay fiber. So, dollar-for-dollar, it gives you a faster bang for your buck.

S-A's open-standards, eight-channel DWDM platform, says Johnson, will multiplex up to eight Prisma Digital Transport systems on the same fiber, increasing per-fiber transmission to 128 video channels. This represents an aggregate data transmission rate of nearly 20 gigabits per second. When used with S-A's Prisma fiber amplifiers, the platform will enable transmission distances of up to 160 km (approximately 100 miles) without the need for Sonet regenerators.

Ortel's new DFB transmitter board

One vendor's perception of the DWDM market was reflected in what it didn't bring to the Western Show. According to Ken Regnier, IPITEK 's (Integrated Photonic Technology) director of marketing and strategic planning, the company decided to leave its 8-channel DWDM solution home in 1997. "We actually showed an 8-channel DWDM at last year's show (1996)," says Regnier. "It was a prototype. But we really just put a floater out there to say, 'OK, where does this stand in our pecking order of product development? Is this something that has such an overwhelming demand that we put everything else aside and get that finished?'

"We gained a lot of experience by showing it. I think the demand will be there, we just don't see it now. But we're poised. So, if customers really start to show (an interest) and are willing to pay the premium, then we'll be there."

In the meantime, IPITEK continues to develop new products in its uncompressed digital, fiber optic transmission and broadband RF analog transmission lines. Some of its newest product additions include a new 1310 DFB laser transmitter module and a 1310 laser transmitter. The transmitter module has a bandwidth of 862 MHz and features a dual RF input with greater than 50 dB isolation, switchable automatic gain control (AGC) circuitry and field adjustment of the optical modulation index (OMI).

A new entrant in DWDM technology is taking a decidedly short-range view of the market. While many DWDM solutions are aimed at long-haul transport, California-based Osicom Technologies has positioned its GigaMux DWDM Fiber Optic Transmission Systems for short-haul metropolitan applications. According to James True, vice president for marketing at Osicom, the company has achieved cost savings for its short-haul (5 to 120 km) 16-channel DWDM system by designing a scalable system from the inside out.

"We designed growth in GigaMux from the core out," says True. "You don't have to buy a complete core on day one. A lot of the products require buying a complete capacity core and then adding channels as needed. What we've done is designed a core that is scalable based upon what the demand is.

"We're also completely transparent to the data format and the speed. And because the metropolitan arena does not require the amplification of long-haul applications, we've removed the cost of that as well. In addition, the whole core of the GigaMux is passive. There's absolutely no power going into it. That makes it immune to power fluctuations and such things as temperature instabilities. Bottom line, the passive design results in an unprecedented meantime between failures (MTBF)."

True says the product is currently undergoing trials at NASA by Dynacs Engineering, a technical services subcontractor at the Kennedy Space Center in Florida. The trial involves carrying critical digital video signals from the launch pad to the Launch Control Center over fiber optic cable.

I'll see your 1310, and raise you 1550

Meanwhile, in the 1550 nm realm, the Broadband Communications Division of ADC has introduced two new Homeworx transmission systems, which include models for single and dual fiber supertrunking, and single fiber distribution.

Ongoing consolidation of systems, with increasing reliance on the creation of superheadends for program distribution, is one factor creating operator demand for 1550 nm gear, says Frank Weiss, product line manager, Homeworx AM Products. "And, to a lesser degree today, but developing more into a trend in the future, is the use of 1550 superdistribution types of architectures, where the broadcast portion of the signal is put on to the 1550 laser at the regional headend, and then transported in fiber directly to the optical nodes. So you have the mother of all transmitters up at the headend, and at the hub locations, you just have EDFAs that boost the signal and send it on its way. That's a very economical way of sending lots of bandwidth to lots of people, from a central location." Shipments of ADC's new 1550 transmission products will begin this April, adds Weiss.

Philip's Diamond Marquise amplifier

Philips Broadband Networks is another firm believer in 1550 nm technology. Tom Tucker, product manager for the company's headend optics, thinks it fills a natural void in the progression of technology in the industry. "One of the driving forces in 1550," says Tucker, "is that over the last few years there (have developed) applications beyond the reach of 1310, and the alternative was digital. And because digital transport is more expensive, 1550 helps create an intermediate cost alternative." Tucker likens 1550 to something "between what 1310 couldn't do and what digital could do."

Toward that end, Philips introduced its new Diamond Link 1550 nm Broadband Transport System as a long-haul trunking alternative to digital transport applications. Several transmitter models and a variety of optical amplifiers are available so that operators can easily accommodate long-range trunking or high-count splitting applications.

Philips is also getting its digital house in order. The Western Show saw the debut of its new single-channel digital system for cable headend interconnects. As such, the system has been designed for applications in which signals in the channel line-up gather at the master headend and are then distributed digitally to various remote headends.

In addition, says John Decker, product manager for long-haul fiber optics, Philips' recent strategic alliance with Artel Video Systems Inc. will soon bear digital fruit. Decker says Philips is leveraging some of Artel's core technology and "then doing a collaborative development that will come up with a new system that is quite a departure from what the industry has seen up to this point, in that it is significantly smaller, has more range and lower cost than other digital interconnect systems available."

Toward more painless upgrades

Aiming to appeal to operators who are hedging their bets about future services and necessary upgrades, several manufacturers have announced fiber products which are designed to be upgraded on the fly.

C-COR Electronics has introduced its Navicor line of AM fiber optic nodes and upgrade capabilities, which features a number of common modules that can be swapped in and out of standard housings. "This set of products, and the architecture that we have come up with, allows cable operators to navigate the upgrade of the broadband network as it evolves into the future," says C-COR President and CEO Scott Chandler.

"Say an operator is doing 1,500 homes per node today. With this architecture, as his network evolves, he can turn some of his amplifiers into nodes, and maybe move it down to 500 homes per node. And if, in five years, a hot spot develops, . . . the operator can surgically go in and make that a 100-home node area."

The Navicor line features the new Quadrant, a four-active output bridger and node which is scalable, with the ability to handle up to four optical transmitters and receivers, and featuring redundant power supplies. The Quadrant is available in both 750 MHz and 862 MHz versions and offers multiple reverse transmitter options.

"The modularity gives operators more flexibility," explains Chandler, "so they don't have to overengineer their networks today. They don't have to say, 'I want three data receivers.' Why three? 'Well, my marketing folks say that there may be really heavy data usage.' The modularity gives the operator the ability, where he can configure it with less today, and if the marketing folks appear right, the network can be upgraded later on."

Harmonic Lightwaves also got into the game with its PWRBlazer Scaleable Node, which rounds out its line of node receivers to include a mini node, suited to fiber-rich architectures, and a "middle-of-the-road" node which falls between the other two in terms of flexibility.

"We came up with a node that can be economically installed , and then could be upgraded easily in the future to handle just about anything the industry could evolve into, whether those services include Internet access, video-on-demand or other interactive services," says Eric Schweitzer, product manager of receiver systems for Harmonic. "Rebuilds are very expensive. And if an operator can put in a node today that has this flexibility, they can avoid having to do a rebuild in the future, which should save money in the long run."

Designers of the scalable node split the RF amplifier into two output modules, which allows the user to configure the node for different numbers of outputs. There are also two extra ports which can be used either for direct powering or direct return input to a dedicated return transmitter.

"Direct return input is starting to become significant, particularly in PCS systems, where you want to keep the cable return and the PCS return on completely different fibers," says Schweitzer, "so that ingress on the cable TV system doesn't wipe out the PCS communications." The node also contains eight optical module slots.

"The other thing we found people were really asking for is redundant powering," says Schweitzer. In response, Harmonic designed the node with two power supplies which operate in redundant modes, and which can be driven by either one of two power buses.

But redundancy is not worth much, says Schweitzer, if the node is not monitored. "Unless you monitor the node, you are not preventing outages, you are delaying them, so we have integrated full network management capability."

And finally, ADC's Broadband Communications Division has also stepped up to the plate to answer MSO requests for scalable optical nodes with its ISX3000 series, a new family of optical distribution nodes.

"One of the key things that we are trying to provide to operators is flexibility, and that is via modularity," says Steve Doherty, product line manager for outside plant equipment at ADC Broadband. "They want to be able to deploy something now, and then, they can change it to (fit) whatever their needs turn out to be. So we have made the enclosure that the electronics are contained within, which we call the casting or the housing, into our platform. And now, we are developing modules that operators can plug together (within the housing) to implement new architectures, to pass new services."

By the end of the year, ADC may be offering additional modules to support WDM, as well as EDFAs.

Higher level of integration

The move toward longer fiber links is also influencing the design of products at the component level. Ortel Corp. has added a couple of new members to its Platinum Performance Series of cable TV products, including a DFB transmitter board and a DFB laser module. The impetus behind the new designs came from a couple of different directions, says Bob Jordan, vice president and business manager of broadband communications products for Ortel.

First, operators require improved performance specifications; and second, power requirements continue to increase, as part of an MSO move toward going longer distances with fiber. "1310 nm laser transmitters generally are good out to about 30 km for the longest link, and after that, 1550 nm transmitters will pick up, but of course, with a big price jump," says Jordan. "For the cases that fit somewhere in that 30–40 km band, the systems will require both increased power and improved linearity."

The company's Platinum Plus series consists of a laser, plus a family of different boards that range from fairly straightforward, featuring simple predistortion, all the way out to integrated boards that feature hybrid amplifiers and some other associated electronics, says Jordan.

The new transmitter boards also include a higher level of integration than ever before, yet feature reduced circuit size and part count, which means that they also use less real estate — the smallest board is about the size of a credit card. The board also features power options between 2 mW and 16 mW, with improved distortion.

As for the new laser module, it differs from its predecessor in that it offers a standard, OC-48 pinout (vs. a proprietary pinout), as well as a negative bias.

Return path insurance?

To give network designers more options on the return path, General Instrument has developed a Frequency Stacking System (FSS) that it believes offers operators "a good insurance policy" against a jammed return path.


David Grubb, GI's vice president of marketing for transmission network systems, says the system essentially quadruples return capacity. "The system takes each of the four return inputs to the node and block converts them to their own frequency band. So, the return from port #1 may go between 100–135 MHz, then the next one goes at another block, etc. That allows the operator to send them all back on one fiber to the headend where the process is undone."

"It's really an upgrade path," says Grubb, "to allow you to put in a node, and if down the road, you're doing telephony or a lot of data and you need more return capacity, it's a way to get that without doing any changes to the physical plant design. That way, you can design for a 500-home node, and you can effectively go to a 125-home node size without ripping up your plant."