How was the company started?
Ofidium's technology was developed by two professors from Monash University in Melbourne, Australia: Arthur Lowery previously founded optical simulation software firm Virtual Photonics (now VPISystems); his colleague Jean Armstrong is an expert in wireless systems, particularly OFDM.
OFDM is the dominant modulation scheme in all kinds of communications from ADSL to 4G wireless technologies like Long Term Evolution (LTE) because it is spectrally efficient and highly resistant to imperfections in the transmission channel.
These are exactly the same characteristics that optical system designers seek for future high-capacity, high-speed optical networks. The big question is: if OFDM is so dominant in other communications media, why hasn't it been adopted in optics?
Ofidium's founders came up with some sound reasons why OFDM hadn't caught on in optics, but went on to find solutions to these concerns, which they patented and published about three years ago. "They convinced people that the physics can really work — that was the point at which the company got started," says chief executive Jonathan Lacey.
What's on the product roadmap?
"There are three steps to having a product, and we're on the second step," Lacey told fibresystems.org. "In the first step we did lots of simulations to prove that the technology was possible. In the second step we're building mock-ups of systems in the lab to prove that we can achieve the performance. The third step is the engineering of a product."
The target is to have saleable products in the form of 300-pin transceivers in the first half of 2011. That's about the time that network equipment manufacturers will be making decisions about what equipment they are going to deploy in volume, according to Lacey.
How much funding has the firm received?
Australian VC firm Starfish Ventures invested A$6 million last September. That funding was only announced last week, in what is likely to be one of many 100 Gbit/s announcements in the run up to OFC/NFOEC later this month.
"The funding is for the second step, which gets us to early 2010, then we'll be looking for another round of funding," Lacey adds.
What is OFDM anyway?
Professor Lowery, the start-up's founder and chief technology officer, likes to use a musical analogy to explain OFDM, Lacey tells fibresystems.org. If the ones and zeros of ordinary data transmission are like transmitting a single note, then OFDM is more like transmitting a chord. Since more data is packed into the chord, this results in slower-speed signals that are less prone to dispersion as they travel down the fibre.
The frequencies (subcarriers) in the chord are created electronically by a digital signal processor, before the signal is converted into light, so the data is still transmitted over a single wavelength.
What are the benefits?
Ofidium believes that its approach will have significant advantages over other 100 Gbit/s techniques. As well as being highly tolerant of dispersion, OFDM should be more robust to the optical filtering found in reconfigurable optical add-drop multiplexer (ROADM) networks.
The most popular approach to 100 Gbit/s transport mdash; albeit not the only one — is based on dual-polarization quadrature phase-shift keying (DP-QPSK) with or without coherent detection (see 100G backers divided on best approach).
On an oscilloscope, the spectral shape of such 100 Gbit/s signals looks like a sinc function, with a central frequency and lots of smaller peaks of decreasing amplitude. The problem with this is that there's a lot of power in the edges. When the signal passes through a system with ROADMs, then each filter chops off more and more power.
In contrast, OFDM produces a spectrum where the spectral shape is a neat rectangle made up from hundreds or even thousands of spectra. This shape lends itself to transmission through ROADM networks, which translates into greater reach, and better network scalability.
Lacey also believes that OFDM receivers will be easier to build than their DP-QSPK counterparts. Most developers of 100 Gbit/s solutions are either using coherent detection or have it on the roadmap because it provides more sensitivity, allowing optical impairments like dispersion to be compensated electronically in the receiver. But a key component, the analogue-to-digital converter (ADC), is particularly difficult to build because it needs to take 56 gigasamples per second, and so far only one vendor (Fujitsu Microelectronics Europe) is claiming to have come up with a solution that's sufficiently fast.
Although Ofidium's OFDM technology also needs a coherent receiver, the required sampling rate is roughly half of that required by most 100 Gbit/s schemes.
Other companies tried and failed with optical OFDM. Why does Ofidium think it can succeed?
It's hard to avoid comparisons with Centerpoint and Kestrel, agrees Lacey, referring to a couple of OFDM start-ups that fizzled out in the telecoms downturn. But there are vital differences between the technology in 1999 and now, he contends.
"Back then they didn't have enough digital signal processing to assemble the 'chords', so they had to build a very complicated transmitter. Their only option was to build an RF oscillator for each subcarrier and modulate each one individually.
"We can easily implement thousands of subcarriers. Those guys had only as many subcarriers as they could fit into a box, which meant that their solution was very expensive."
Now that the industry has come to accept that coherent detection will be needed at 100 Gbit/s, OFDM is not such a huge step, and in fact has become a topic of huge research interest. At ECOC last September, about 60 papers were submitted on the topic of OFDM, and there will be even more papers on the subject at this years OFC/NFOEC.