Tunable pluggable XFP is the end game, but XFP-E may still play
by Meghan Fuller Hanna
In late February, a large carrier began circulating an RFQ asking for tunable pluggable optics in the 2009 timeframe, and that RFQ has had something of a ripple effect on the industry. Module vendors who were working on a tunable pluggable XFPâ��the â��Holy Grailâ�� of the marketâ��now may consider developing a tunable pluggable XFP-E in the interim to satisfy sooner-than-expected customer demand. With its larger size and less stringent thermal requirements, the XFP-E would enable module vendors to bring tunable pluggable technology to market faster.
The XFP-E is essentially a doublewide XFP, comprising two XFP transceivers placed side by side and integrated within a single module. The surface area may be larger, but its key advantage is its power budget, which is around 6 W or almost twice that of the thermally constrained XFP. But some industry insiders believe that may be its only advantage.
â��We are of the opinion that there is no reason why you would use something that is twice as big and consumes 50% more power,â�� asserts Craig Iwata, senior director of marketing and business operations at JDSU (www.jdsu.com). â��If given the choice, youâ��re going to go for the increased density that you can get from being half the size.â��
Todd Swanson, vice president of sales and marketing at Finisar (www.finisar.com), agrees. â��In 2009, if I had the ability to design an XFP or an XFP-E with 300-pin-type performance, which would you choose?â�� he muses. â��There is absolutely no advantage to an XFP-E because itâ��s bigger, it takes up more slots, and it consumes more power.â��
The problem is that carriers and system vendors do not yet have a choice; the module vendors are still developing the optics and electronics necessary to enable the requisite functionality in a small form factor that consumes just 4 W of power.
â��The central element here is the 300-pin,â�� explains Swanson. â��The system manufacturers and the carriers are very reluctant to change their performance specifications due to the interoperability challenges that presents. So they are going to drive very hard to their suppliers, â��You need to meet my existing specs.â�� And the existing spec, in this market, is the 300-pin.â��
Herein lay the problem: How do you get all the functionality of the 300-pin into the smaller form factor? The folks at JDSU say they have overcome this technical challenge and are on track to unveil their tunable pluggable XFP at the end of this calendar year, thanks to the development of what the company claims is the industryâ��s smallest transmitter optical subassembly (TOSA).
â��A year ago, the world wasnâ��t sure that anyone could get into an XFP form factor,â�� Iwata recalls. â��I believe we have crossed the main hurdle to get there, which was the TOSA that we announced [at the end of January].â��
While several of its competitors are pursuing a co-packaged technology in which the laser and modulator reside on separate chips, JDSUâ��s TOSA is enabled by its integrated laser Mach-Zehnder (ILMZ) chip, which features a monolithic integration of the laser and modulator. The company says the chip itself is small enough to fit on the tip of a finger.
Because JDSU made the strategic decision to develop the TOSA for tunable pluggable XFPs, it has no intention of developing the XFP-E, which it views as simply an interim step. The end game, JDSU believes, is the smaller-form-factor XFP.
Not everyone agrees with this assumption, however. Chris Clarke, chief engineer at Bookham (www.bookham.com), believes there is a customer base that wants to replace its discrete devices or 300-pin MSA transpondersâ��currently used with their ITU-specified G.709 long-haul interfacesâ��with tunable pluggable transceivers, and â��that will definitely not be an XFP form factor,â�� he says. â��It will probably never be an XFP form factor. Thatâ��s a bold claim,â�� he concedes, â��and I could be proven wrong in the fullness of time, but if you look at what is in an MSA 300-[pin] transponder now, that physical functionality will not fit into an XFP.â��
Clarke believes system vendors will be forced into the XFP-E for long-haul applications because thatâ��s the smallest form factor that will accommodate all the requisite functionality and still dissipate the heat. â��You need to have the full transmit and receive chain,â�� he explains. â��You might have to have electronic equalization in there...and all the wavelength locking circuits, the temperature-tracking circuits just to make sure you donâ��t get any penalties over temperature and over life [of the laser]. Every tenth of a decibel is important in this application,â�� he notes.
Clarke confirms that Bookham will be moving to controlled availability of its XFP-E in the fourth quarter of this year and has as a lead customer a Tier 1 company that he characterizes as â��one of the prime movers and shakers on the XFP-E multisource agreement activity.â�� Moreover, he reports that three or four additional customers have â��U-turned from going down the XFP route to wanting an XFP-E because time-to-deployment is more important to them than getting it into the small form factor, which might take another year or so.â��
Bookham is also developing a tunable pluggable XFP, though itâ��s taking a different route than JDSUâ��s monolithic integration. Bookhamâ��s strategy is to use a different materials technologyâ��aluminum quaternaryâ��to enable its XFPs to run at a higher temperature. â��That solves the thermal problem, but weâ��re at least a year away from being able to do that, maybe a little bit longer,â�� says Clarke.
In fact, the biggest remaining question seems to be one of timing, not technical feasibility. While JDSU maintains that its tunable pluggable XFP will be available at the end of this calendar year, Finisarâ��s Swanson stresses that it may be several yearsâ��likely 2010 or even 2011â��before we see multiple suppliers offering full C-band tunability in a pluggable XFP form factor.
In other words, while the tunable pluggable XFP remains the Holy Grail, to use Iwataâ��s words, there may be life yet in the XFP-Eâ��assuming the carriers are unwilling to wait for the smaller form factor. â��If someone could come up with a pluggable module that is lower cost than a 300-pin, customers would be extremely excited about that,â�� Swanson contends. â��But the challenge, of course, is to do both of those things [pluggability and low cost]. And to do both of those things and put it into an XFP in a short timeframe I donâ��t think is something that is practical for most suppliers.â��
When asked whether Finisar will now develop an XFP-E to satisfy more near-term requirements while it is finalizing its XFP design, Swanson admits that the company is â��investigating that option. You could imagine thatâ��s the prudent thing to do when the dynamics of certain large carriers and customers change,â�� he says in reference to the recently released RFQ.
No doubt other vendors are rethinking their tunable pluggable strategies as well.
Meghan Fuller Hanna is senior editor at Lightwave.