Two industry sources have confirmed that CableLabs has begun work on a data-over-coax project that looks beyond Docsis 3.0 and might bring an end to the practice of dividing bandwidth into 6MHz or 8MHz channels.

The discussions, reported by Multichannel News on Sunday, are in the "pretty early stages," a source confirmed to Light Reading Cable. "This would be a big effort," the person added, saying it would require the development of new chips and new systems.

According to the report, CableLabs and its members are working from scratch, targeting capacities of 5 Gbit/s or more using new media access control (MAC) and physical layer (PHY) components.

That means it would be years before products are developed, and even longer before they might be deployed.

"I assume they're planning for the second half of the decade, because this would be a massive effort," says Heavy Reading senior analyst Alan Breznick, noting that several operators are already looking at ways to graft EPON to their Docsis systems for business services and some limited residential service scenarios.

And no one is suggesting that the industry is getting ready to migrate to a channel-free, and possibly more efficient, spectrum environment anytime soon. "Docsis 3.0 is very strong, and has good legs," notes an industry source who is familiar with the CableLabs project.

CableLabs declined to comment. The Louisville, Colo.-based R&D house told Multichannel News that it "is in business to explore many different technologies on behalf of its members."

Getting to 1 Gig
Capacity of 5 Gbit/s already exists on most upgraded cable plant, but the bandwidth is divided into the aforementioned channels. (North American Docsis uses a 6MHz channel width; EuroDocsis uses 8MHz.) A plant built to 750MHz supports capacity for almost 5 Gbit/s, while 1GHz plants support more than 6 Gbit/s, with the individual channels divided up for the actual cable services (broadcast video, video-on-demand, switched digital video, VoIP, high-speed data, etc.).

And Docsis 3.0 is perfectly capable of delivering burst speeds in excess of 1 Gbit/s. Vendors have already told Light Reading Cable that they are working on D3 modem configurations that can bond eight or 16 downstream channels, and are seeing requests to bond 32 downstream channels -- enough for a 1.2 Gbit/s burst. (See The Ultimate Cable Modem .)

Comcast Corp. and other MSOs have been deploying wideband aggressively; Heavy Reading forecasts that among North American MSOs, Docsis 3.0 will pass 90 million homes by 2012. (See Wideband Growth Continues in North America.)

Comcast, meanwhile, is hooking up with an array of partners on a next-generation access architecture project that mixes Docsis 3.0, PON, and a way to help cable migrate to more IP-delivered video services. (See Cox Adds Weight to Comcast's Big Box Project and Comcast Proposes Its God Box .)

Even if the CableLabs project is just starting, the fact that such work is underway indicates that the cable industry intends to squeeze as much out of its hybrid fiber/coax (HFC) plant as it can before moving to a fiber-to-the-home (FTTH) platform.

That has not stopped operators such as Canada's Shaw Communications Inc. from considering the use of FTTH architectures and Radio Frequency Over Glass (RFoG) in all their new buildouts. (See Shaw Vetting FTTH for All New Builds and Costs Could Keep RFoG a Niche Player .)

The CableLabs effort also calls into question whether the consortium will go to the lengths to specify a Docsis-optimized FTTH platform that uses the technology's management plane. A July 2006 CableLabs report called "Cable Response Alternatives to FTTP" caused a bit of a stir by suggesting that it might be more cost-effective for MSOs to deploy FTTP when node sizes are less than 125 homes passed.

— Jeff Baumgartner, Site Editor, Light Reading Cable