Your Friendly Neighborhood 100 Gbps Modem

B. Washburn

B. Washburn

Summary Bullets:

  • 100 Gbps wavelength deployment appears to be picking up, a year later than expected.
  • A good deal of optical gear is using 25 Gigabaud modem techniques to achieve 100 Gbps.

I have been on the fence about 100 Gbps wavelengths for a while now.  Sure, the technology will become mainstream eventually.  However, 10 Gbps long-haul waves came through the door years ago and ramped up fast; they are cheap and plentiful on competitive routes.  Meanwhile, major U.S. service providers touted that they would add 100 Gigs commercially by late 2010, but well into 2011, the U.S. service providers as a whole had only confirmed operating a handful of 100 Gbps terrestrial routes internally.

Another reason I have not been enthusiastic about 100 Gigs is that some early products to market reached at first 40 Gbps, and now 100 Gbps, by stapling together a bunch of lower-speed (10 Gbps) waves.  A similar thing happened back in 2001, when some optical vendors lashed together 1 Gbps wavelengths to hit 10 Gbps; that technique was, and is, a stop-gap measure that does not change optical economics and drive the industry forward.

These factors look like they are now starting to change.  Carrier adoption of long-haul 100 Gbps deployments seems to be picking up some steam, using coherent signaling – that is, using one signal instead of lashing together separate, lower-speed wavelengths.  Furthermore, the industry has fundamentally shifted transport techniques to achieve long-haul 100 Gbps.  Many coherent, long-haul transport devices have adopted variations of quadrature phase shift keying (QPSK) for optical transport, a technique that stuffs more data onto a signal by encoding it as bit pairs.

To explain, each individual transported bit is a binary zero or one; QPSK divides transport into bit pairs (called symbols) that are read together:  Two bits turn into four possible states, doubling data throughput.  So far, an optical industry favorite QPSK flavor is differential polarization (DP-QPSK), which coordinates two lasers on the same frequency.  Inside the optical transport box, a 100 Gbps stream becomes two 25 Gbps carriers transmitting 25 Gigabaud (symbols/second); at the far end, another box re-constitutes the stream of symbols back into a single 100 Gbps stream.

QPSK might be new in optical transport, but it is old hat for satellite terminals, dialup modems, microwave, and broadband.  QPSK and its more complex cousins, three-bit/eight-phase 8PSK and four-bit/16-phase 16QAM (quadrature amplitude modulation), were designed to stretch extremely limited spectrum.  Getting silicon to disassemble and re-assemble a 100 Gbps stream as dual 25 Gbps streams of symbols is a remarkable achievement, though it seems odd to use terms like “baud” when discussing optical wavelength speeds.  DP-QPSK is not the only way to get to 100 Gbps; it may even get eclipsed by other modulation techniques in the long run.  However, it is great to see old-fashioned technologies developed for the humble data modem find new life at the industry’s cutting edge.

About Brian Washburn
Brian Washburn is Research Director for Network Services at Current Analysis. Brian tracks the technology and initiatives surrounding carrier Ethernet, IP-VPNs, optical networking and applications closely tied to high-performance networking.

What do you think?

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: