Vendors weigh tunable pluggable challenge
by Meghan Fuller
Today, tunability is available in the 300 pin multisource agreement (MSA) transponder and is used mostly for sparing and inventory management in long-haul and ultralong-haul DWDM applications. Small-form-factor pluggables (SFPs), by contrast, are primarily used in low-cost, short-reach applications, also for sparing and inventory management purposes. But what if you could somehow combine tunability and pluggability in the same small form factor? Would you reap additional benefits? For many industry insiders, the answer is a resounding “yes.”
Pluggable transceivers or SFPs were first used in the telecom market for CWDM applications where the number of wavelengths were limited to, say, four or eight, recalls Pierluigi Franco, senior vice president of marketing and product management in Pirelli Broadband Solutions’ (www.pirelli.com) Photonics Business Unit. In this case, tunability would not provide additional benefits; using pluggable optics, network operators already have reduced the number of spares they have to carry.
“But now pluggable transceivers are becoming DWDM, and that means the number of wavelengths will increase again,” reports Franco. “You need both pluggability and tunability to reduce the number of spare transceivers. And this, I believe, is the reason why the combination of tunable and pluggable is something that most everybody is looking for,” he asserts.
Chris Coldren, business development manager at JDSU (www.jdsu.com), agrees that the combination of tunability and pluggability “makes complete sense” in the emerging DWDM pluggable market. He cites the 10 Gbit/s DWDM XFP as an example. “We’re getting feedback from customers of those products, saying, ‘Hey, DWDM XFP is great, but the same problems we had with fixed-wavelength transceivers are coming to hurt us again. These things are expensive and not that easy to get a hold of. We don’t want to have to inventory hundreds and hundreds of these things when one part code should do, if it were tunable’.”
Moreover, tunable pluggable transceivers would further reduce inventory by enabling network operators to limit their spare inventory to the optics, not the line cards, says Franco. “I hear people say, ‘If it is already tunable, why do I need it pluggable? If I have one board that is tunable and can replace any other board, I probably don’t need pluggable optics.’ This is something that I really do not agree with,” Franco asserts, “because the pluggability gives you some advantages.”
The combination of tunability and pluggability would enable individual boards or line cards to be reconfigured or repurposed simply by populating them with different pluggable modules. Today, tunable lasers are hardwired onto the boards, so although network operators only need one laser to cover the entire C-band, for example, they still need to manufacture boards for specific applications. But if they had tunable pluggables, they would be able to reconfigure boards for any application by adding or removing tunable pluggable devices.
On paper, tunable pluggables seem like a no-brainer. In reality, they represent a formidable technical challenge: How do you shrink a tunable transmitter into a small form factor while adhering to the MSA power dissipation requirements? Standards for existing MSA form factors-XENPAK, XPAK, X2, and XFP, for example-were designed to accommodate fixed-wavelength lasers and have limited margins for size and power dissipation.
The current generation of tunable modules consumes around 11 W maximum, says Ed Cornejo, director of product marketing at Opnext (www.opnext.com). “That’s a lot to put into one of these smaller, pluggable form factors. We’ve specified X2s with a maximum power dissipation or thermal capacity of 4 W, so you can see there’s a lot of disparity between 4 W and the current [tunable] technology of around 11 W,” he notes.
The majority of today’s tunable lasers cannot be directly modulated; they require an additional modulator outside the laser cavity, which requires yet more real estate and consumes additional power.
That said, most component vendors seem undeterred by the challenge of combining tunability and pluggability in a small form factor. “From JDSU’s standpoint, we’re saying, ‘Hey, this is an important segment because, in many ways, it’s an additional market beyond the business we’re already doing in the long haul and ultralong haul,’” confirms Coldren. He notes that JDSU has already demonstrated a monolithic wideband tunable laser with integrated Mach-Zehnder modulator-on a single chip. “What we’re working on now is getting that into the transceiver form factor,” he says.
According to Adam Price, director of product line management for Bookham’s (www.bookham.com) Mach-Zehnder and tunable transmitter portfolio, Bookham also has what he says are the two key components required to make a tunable pluggable device: a credible laser technology and a credible modulator technology. At last year’s OFC/NFOEC, the company demonstrated a wideband indium phosphide (InP)-based tunable transmitter assembly (TTA) module and integrated InP Mach-Zehnder modulator, says Price.
Franco reports that Pirelli also is making progress on a tunable pluggable. “At ECOC last September, we announced a 200 km 10 Gbit DWDM XFP,” he says, “and, of course, we have a tunable product line with a tunable iTLA and tunable MSA 300 pin. The technical challenge, obviously, is to squeeze everything into the small form factor, but I believe it is feasible”.
Opnext’s Cornejo reveals that his company is “trying to create new tunable technology,” but he is not so sure that pluggables will feature full-band tunability in their first incarnations. “Even with some of the monolithic designs that are out there, you are still going to be fairly high power,” he admits.
For this reason, he says, the first tunable pluggables may be only narrowly tunable over, say, four or eight channels depending on the spacing. And if you have a module that doesn’t cover the full bandwidth, pluggability really begins to make sense, he explains. “If you had a situation where you needed a channel that was on the other end of the spectrum, that one module wouldn’t be able to tune to that wavelength; you’d have to plug in a different module,” he says, “and that would be one reason to have a pluggable.”
There is some precedence for incorporating narrowband tunability; the first tunable lasers were narrowly tunable distributed feedback (DFB) lasers thermally tuned over a few wavelengths. However, not everyone is convinced that tunable pluggables will follow this trend.
“JDSU’s position is that’s kind of a band-aid,” contends Coldren. “It’s not the fix. The fix is really going full-band tunable.”
Franco agrees. “Looking at the technology that I see around and also inside Pirelli,” he says, “I believe that either people will target full tunability from the beginning or limited tunability but something like 10 nm to 12 nm, which is more than what is possible by simple thermal tuning.”
The industry also continues to debate which pluggable form factor would most benefit from tunability, though the early frontrunner appears to be the XFP. “For Pirelli, pluggable tunable means XFP,” asserts Franco, who is convinced that such prototypes and/or demonstrations will occur within this calendar year.
When asked whether a tunable pluggable XFP would impinge on the market share currently held by 300 pin transponders, Franco admits that it could, but he also believes the 300 pin will carve out a new niche. He speculates that tunable XFPs could dominate the lower-end market, say, 120 to 200 km. The 300 pin MSA transponder, with integrated dispersion compensation techniques, could capture the very long reach, 350 to 400 km market-not to mention the 40 Gbit/s market.
“If you ask me what I see in three, four, five years-and I’m really guessing-but a possible situation is 2.5 Gbit SFP, 10 Gbit XFP, and 40 Gbit MSA,” he says.
For his part, Bookham’s Price would not commit to his company’s specific plans, though he did say that if the market embraced the XFP footprint, “in terms of optics, we are looking at enabling ourselves in that type of footprint space, if necessary.” Price does confirm that Bookham has a tunable pluggable on its roadmap, though he doesn’t think the device will come to market until 2008. “It’s all subject to market demand in terms of urgency,” he says.
While Opnext’s Cornejo also admits that a tunable XFP is the ultimate goal, he believes such a device is “a pretty long way from now.” He’s not entirely sure Opnext’s customers need a tunable pluggable today. “It’s been an interesting discussion we’ve had with customers,” he admits. “Some of them would be very happy just having a small-form-factor 300 pin, even like a 2 inch by 3 inch tunable. That would get them to some of the densities they are looking for.”
Nevertheless, Cornejo reports that Opnext is doing some background work on a tunable XFP-E, a doublewide XFP transceiver based on an MSA that the company developed with Avanex and Mitsubishi Electric. He says Opnext is also looking at X2 as a potential form factor (presumably because Cisco, the main proponent of the X2, is a key Opnext customer.)
Per Hansen, senior director of business development of the Enterprise Optical Division of ADVA Optical Networking (www.advaoptical.com), believes that system vendors will embrace tunable pluggable transceivers such that “you may see pressure toward this just being table stakes.”
For benefits, he cites the popular inventory reduction argument first. But Hansen also sees potential benefits from another perspective: adding pluggability to high-end tunable transceivers.
“From a system vendor’s perspective, one of the great advantages will be that you will get a standardised interface,” he says. “One of the great benefits you’ve gotten out of pluggables-independent of their performance-is that there are a lot of vendors who all build to nearly the same spec. To some degree, they are interchangeable, and that just gives you a higher competitive situation in the market. You’re likely to see the world push for more uniformity and commodity pricing, if you will, on those devices,” he adds.
In terms of the form factor debate, Hansen notes that space is at a premium in systems today. “Because you have so many interfaces on a transponder card, there will be a strong preference among system vendors to go toward smaller form factors,” he says, adding that, “We often compete on density.”
Moreover, he says, “I would expect that the market and the system vendors would really be interested in the full-band tunable to harvest all the advantages.” He confirms that ADVA Optical Networking is “certainly going to be interested in utilising that as well.”