Ofidium Tackles 100Gig
Brings back memories of Centerpoint and Kestrel, doesn't it? (See our 2001 Optical Taxonomy and Metro DWDM.)
Those were 10-Gbit/s ventures, though. Australian startup Ofidium is trying its hand at OFDM for 100-Gbit/s optical networking.
Barely a year old and staffed by an unspecified handful of employees (CEO Jonathan Lacey will only say the number is really small), Ofidium was founded by Monash University professor Arthur Lowery -- who'd also founded VPIphotonics, which makes software for optical-components design. Ofidium was lucky enough to land $6 million in funding from Starfish Ventures Pty. Ltd. in September, just before the world economy tanked.
That funding round got announced just today. (See Ofidium Scores $6M.) It's the first of what will likely be many, many 100-Gbit/s announcements preceding this month's OFC/NFOEC show.
OFDM is the same concept powering 4G wireless technologies such as Long Term Evolution (LTE) and mobile WiMax, so it's not that radical an idea. The physics of Ofidium's approach works; it's a matter of proving that the engineering is feasible, which is how Ofidium will be spending that $6 million.
The company is hoping to release prototype transceivers, which won't be suitable for commercial sales, in the first quarter of 2010. That will give equipment vendors time to play with the technology while Ofidium prepares its real devices for early 2011 shipments.
That would put Ofidium on schedule to have parts ready when the Institute of Electrical and Electronics Engineers Inc. (IEEE) 802.3ba standard for high-speed Ethernet gets ratified.
Ofidium's transceivers will use OFDM to split a 100-Gbit/s flow into hundreds or even 1,000 lower-rate streams, all transmitted along the same optical wavelength. The splitting is done in the electrical domain -- before the signal gets converted to light -- resulting in slower-speed individual signals that are much less prone to dispersion as they travel down a fiber.
Other optical companies are talking about transporting 40- and, ultimately, 100-Gbit/s signals by dividing them into multiple signals. The common approach, pioneered by Nortel Networks Ltd. , has the slower signals traveling on multiple "carriers" within one ITU-grid wavelength -- meaning, the transmission looks like a single wavelength to the receiving equipment.
Companies pursuing variations on that approach include Ekinops SA , Mintera Corp. , OpVista Inc. , and StrataLight (now part of Opnext Inc. (Nasdaq: OPXT)). (See Ekinops Pushes 40-Gig, OpVista Runs With DMC for 40-Gig, and Nortel Rolls On With 40-Gig.)
On an oscilloscope, the spectrum for those kinds of transmissions looks like a sync function (think bell curve with oscillating tails), while the OFDM method produces a rectangular shape built of hundreds of little spectra packed together.
What's the big deal? It comes down to dispersion, the distortion that high-speed optical signals undergo during transit. More dispersion means more signal processing is required at the receiving end. A chip called an analog/digital converter (ADC) is one of the weak links there, because most 100-Gbit/s schemes will require an ADC that can do 56 billion samples per second, something only Fujitsu Microelectronics Europe (FME) is claiming to have accomplished. (See Fujitsu Micro Tackles 100G.)
Ofidium can live with an ADC that's less than half as fast -- which is something that's more readily available.
"That's one of the biggest attractions," chief executive Lacey says. "They know that the components that we need to build this approach are available now or will be shortly."
One hitch is that neither Centerpoint nor Kestrel made it as startups, giving OFDM a less than stellar history in optical networking.
"I was at Verizon Communications Inc. (NYSE: VZ) then, and we had invested heavily into WDM. The question was what advantage we would gain by changing from WDM over to this 'FDM. It really didn't help us at all," says Sam Greenholtz, a principal analyst with Telecom Pragmatics Inc.
Then again, Greenholtz and fellow principal Mark Lutkowitz agree that the shift to 100 Gbit/s speeds is radical enough that new technologies could have a shot.
"I've talked to the other companies at the high end, and they're not snickering at 'FDM," Lutkowitz says. "When you're dealing with 100 Gbit/s, the sky's the limit."
Greenholtz's skeptical side took over, though, once he was told Ofidium didn't exactly have products yet. "Vaporware never fails," he said.
— Craig Matsumoto, West Coast Editor, Light Reading