Cable-Tec Expo: DOCSIS 3.1 to Blaze Trail Toward 10 Gig SpeedsCableLabs: Spec Won’t Require Plant Upgrades, Will Be Backward-Compatible With 3.0 10/18/2012 9:14 AM Eastern
Orlando -- CableLabs and MSO executives provided a detail-rich discussion of the industry’s plans for DOCSIS 3.1, the next generation of cable-modem technology that promises to provide data speeds of 10 Gigabits per second or more.
The DOCSIS 3.1 spec uses new modulation technology and eliminates the traditional 6 MHz channel divisions, to use spectrum more efficiently. However, 3.1 will require backward-compatibility with existing 3.0 cable modems and CMTSs, and will not force MSOs to make any changes on their HFC plant to deploy initially.
“I can’t emphasize this enough: you don’t have to change you plant to deploy DOCSIS 3.1,” Matt Schmitt, CableLabs director of DOCSIS specifications, said. “It can maximize what you can get out of your existing plant, but make it future-proof.”
Schmitt spoke at a 90-minute special session jointly presented by the Society of Cable Telecommunications Engineers and CableLabs here Thursday.
CableLabs expects to publish the DOCSIS 3.1 spec in 2013, and hopes to enable equipment vendors to deliver their first products in 2014.
“It’s a pretty quick follow-on to 3.0 and is a very substantial change,” SNL Kagan analyst Ian Olgeirson said. “I foresee challenges ahead for the development of the spec and the equipment, and the eventual implementation.”
The 3.1 spec will be designed to efficiently provide 10 Gbps of downstream and at least 1 Gbps upstream, “probably more,” Schmitt said. “That’s where we believe we’ll need to go with the technology.”
At the same time, CableLabs is aiming for a significant cost reduction in terms of cost per bit delivered with DOCSIS 3.1. “It should be dropping significantly from one generation to the next, and that’s absolutely a key component of what we’re trying to do,” Schmitt said.
DOCSIS 3.1 uses a new forward error correction (FEC) scheme, Low Density Parity Check (LDPC). That will let MSOs move to higher-order modulations, from 256-QAM today to up to 4096-QAM or even beyond, according to Schmitt. The advantage: 4096-QAM is 50% more efficient than 256-QAM.
“You get more bits per second over the same spectrum,” he said.
For example, with DOCSIS 3.1, delivering 5 Gbps of downstream capacity will take 500 MHz of spectrum, whereas 3.0 would take 780 MHz.
To gain even more efficiencies, DOCSIS 3.1 also incorporates orthogonal frequency division multiplexing (OFDM) for both upstream and downstream. That will create 20-50 KHz channels transferred in a block – "mini-channels” that are in a sense bonded together -- to break the traditional 6 MHz channelization (or 8 MHz in Europe).
Another benefit of OFDM is that it’s widely adopted, and the cable industry hopes to “ultimately attract a larger ecosystem of chip set vendors into the ecosystem,” Schmitt said.
DOCSIS 3.1 will be required to be backward-compatible with 3.0. On the headend side, a 3.1 CMTS or Converged Cable Access Platform (CCAP) device will have to support at least 24 downstream and eight upstream DOCSIS 3.0 QAMs, Schmitt said.
That hybrid 3.0/3.1 approach will let MSOs start buying DOCSIS 3.1 modems or gateways even if they don’t have headend equipment yet.
Comcast vice president of access architecture Jorge Salinger said he expects the MSO to deploy 3.1-based customer premises equipment first, then turn on DOCSIS 3.1 in the downstream before finally -- if necessary -- switching on 3.1 for the upstream.
“We wanted to develop a smooth transition strategy,” he said.
Salinger said he expects the cost difference between 3.0 and 3.1 modems to be “significantly smaller” than the price delta the industry saw between 2.0 and 3.0 CPE. “Things won’t change as much in a big way,” from DOCSIS 3.0 to 3.1, he said.
In terms of spectrum allocation, DOCSIS 3.1 will maintain upstream data below the downstream side (some in the industry had proposed a "top-split" with upstream placed in higher frequencies). With 3.0, the upstream resides in the 5-42 MHz band; the exact number for 3.1 is to be determined but will be around 200 MHz. On the downstream side, 3.1 will operate up to 1.1 or 1.2 GHz, Schmitt said.
Howard Pfeffer, Time Warner Cable senior vice president of broadband engineering and technology, called DOCSIS 3.1 “an evolutionary step along the way of finding spectral efficiency.” While LDPC is more computationally expensive than other FECs, DOCSIS 3.1 will be able to take advantage of Moore’s Law for faster processors.
Also important is that DOCSIS 3.1 preserves the “DOCSIS service plane,” Pfeffer said, such as scheduling, security, filtering, IP source error correction and fraud prevention.
The SCTE wants to help the industry from the get-go on DOCSIS 3.1, so it can be deployed as quickly as needed, said Daniel Howard, SCTE’s CTO and senior vice president of engineering. As a key step, the group has formed a Special Working Group within the SCTE Standards Program that will develop best practices and requirements to prepare HFC networks for DOCSIS 3.1 and other higher-capacity signaling schemes.
The goal was to determine, “How do we start planning for this yesterday so that it rolls out very quickly?” Howard said.
“In the past, it’s been serial,” with CableLabs developing a specification before SCTE worked up training programs and certification. “We can’t take that time anymore.”
CableLabs over the last two years has done “pre-work” on the DOCSIS 3.1 specification, vice president of access network technologies Dan Rice said. In June, it brought in vendors to start the specification project in earnest. Then, in July the consortium settled on key technology tenets with MSO members as well as established intellectual property agreements.
“We don’t have to debate a lot of the basic technology decisions,” Rice said.
Rice, in his prefatory remarks, patted the cable industry on the back, saying that it “has done an excellent job of hitting the exact right spots in helping the Internet evolve… Our technology is just ahead of where it needs to be.”
In the past decade, operators have generated about $230 billion in revenue from DOCSIS residential services in the U.S., he noted.
DOCSIS 3.1 incorporates several of the key recommendations in “DOCSIS NG” (for “next generation”), a white paper presented at the 2012 Cable Show in May by Arris Group, Cisco Systems, Motorola Mobility and Intel. Those include the adoption of OFDM and LDPC, and coexistence with existing DOCSIS 3.0 equipment.