The Docsis 3.1 specs under development at CableLabs won't be published until later this year, but it's already well-established that cable's latest, greatest data platform will make big use of
orthogonal frequency division multiplexing (OFDM) as the industry begins to break away from the use of less efficient 6MHz- and 8MHz-wide channels. (See Docsis 3.1 Set for a Spring Fling.)
OFDM will provide Docsis with a new order of bandwidth efficiency, but it also offers some sound business reasons for cable to adopt it. OFDM, already used for Wi-Fi and Long Term Evolution (LTE), could lead to better economies of scale and get more vendors interested in the cable market, explained Daniel Howard, the SVP of engineer and CTO at Society of Cable Telecommunications Engineers (SCTE), during a webcast earlier this month on the topic (SCTE is tasked with helping the cable industry get trained up for Docsis 3.1). (See Setting the Stage for Docsis 3.1.)
But back to the coming efficiencies. By letting MSOs scrunch small (20KHz-to-50KHz-wide) subcarriers into blocks of spectrum (or "waveforms" up to about 200MHz wide) without interfering with each other, OFDM essentially eliminates the wasteful guard bands required for today's single-channel QAM technology.
OFDM also deals well in noisy environments and can handle other impairments such as micro-reflections, meaning cable can be better able to maintain higher forms of modulations and push more bits per hertz on the cable plant even when plant conditions aren't ideal. OFDM's use of adaptive bit loading also allows operators to dial those modulations up and down at the subcarrier level. In the SC-QAM world, the entire 6MHz or 8MHz channel must operate at the same modulation.
Howard also pointed out a popular misconception that OFDM alone will help operators achieve higher modulations that can put more bits down the pipe in the same amount of space. (See Docsis 3.1 Targets 10-Gig Downstream.)
To hit those anticipated, improved performance levels, OFDM will be paired with low density parity-check (LDPC), a Forward Error Correction (FEC) scheme that will also require less bandwidth than the current Reed-Solomon approach. When used together in Docsis 3.1, they should improve cable's bandwidth efficiency by about 50 percent. So it follows that cable should be able to pump out 5Gbit/s in 500MHz of spectrum with Docsis 3.1, versus 780MHz of spectrum using today's Docsis 3.0 platform.
Although Docsis 3.1 will help cable push the limits and extend the life of the hybrid fiber/coax (HFC) plant, the irony is that the industry will owe a debt of gratitude in part to a technology that's helped wireless operators get the most out of their precious spectrum.
— Jeff Baumgartner, Site Editor, Light Reading Cable