Acacia Takes a Stab at 100G

Acacia Communications Inc. bills itself as a 100Gbit/s startup, but it's possible the company already has its sights set on higher speeds.

Officials at the Maynard, Mass.-based startup expect to launch a 100Gbit/s product this year, as they told the Boston Business Journal in November. The product sounds like it's a module or subsystem, and it's supposedly in beta tests with equipment providers.

But there's also this: Acacia lured photonic integration expert Christopher Doerr away from Bell Labs this fall. And that might be a sign that Acacia is working on 400Gbit/s or faster speeds, says Sterling Perrin, a Heavy Reading analyst. (Perrin has no actual knowledge of what Acacia is up to -- he's just humoring us here.)

"This is pure speculation, but I wonder if they're starting to look beyond 100Gbit/s," Perrin says. "A new company could get a foothold there because it's just now starting to get going, and something like photonic integration could play a huge role in this next phase, whereas the role for photonic integration at 100Gbit/s could be limited."

Acacia was founded by Benny Mikkelsen, who also founded 10Gbit/s and 40Gbit/s startup Mintera, and it's got Eric Swanson, a founder of Sycamore Networks Inc. (Nasdaq: SCMR), on its side. (See OFC/NFOEC: Startups Chase 100G.)

The world doesn't know much else, not even how much Acacia has raised from investors Matrix Partners , Commonwealth Capital Ventures and OFS in two rounds.

Acacia has turned down most interview requests so far, including Light Reading's. (Sad face.)

Doerr is definitely there, though. Sources say so. His LinkedIn profile (registration required) says so. And the bio for this course he's giving at OFC/NFOEC in March says so.

Why does that suggest a focus on post-100Gbit/s technology? Because most of the magic for 100Gbit/s coherent links lies in the electronics, not the optics. Systems vendors have been applying digital-processing skills to create their own ASICs for 100Gbit/s modules. (See Fujitsu Readies 100G Optics and Juniper Amasses 100G Optical Team.)

But for higher-speed modules, whether they're 400Gbit/s or 1Tbit/s, photonic integration might provide some big advantages.

"Chris Doerr was always a future-looking guy. I could see him being interested in that, rather than a rear-view-mirror product, since the industry has already defined 100Gbit/s," Perrin says.

There's also the fact that the 100Gbit/s market is going to be hotly competitive.

"The modules business is rough. A couple of big companies dominate, and there's not going to be a lot of room," Perrin says.

And he's not sensing much desire among systems vendors to have more than a couple of 100Gbit/s module suppliers on the market, which doesn't bode well for a startup trying to squeeze in.

Optical startups have become rare in telecom. Ofidium Pty Ltd. was working on 100Gbit/s modules, with thoughts of applying orthogonal frequency division multiplexing (OFDM) to the problem, but the industry has instead adopted the Optical Internetworking Forum (OIF) framework, which is based on coherent detection. Ofidium shut its doors in July, according to former CEO Jonathan Lacey's LinkedIn profile (registration required).

— Craig Matsumoto, Managing Editor, Light Reading