Broadcom Corp. engineers are working on a way to bring PON speeds to cable so MSOs can run symmetrical 1Gbit/s speeds, extending the life of hybrid fiber/coax (HFC) networks.

It's all still in the proof-of-concept stage, but the chipmaker is hopeful that its approach -- something it's calling EPON-over-Coax (EPoC), a way to graft the coax PHY layer to the EPON MAC layer -- will eventually evolve into a standard. Broadcom has submitted a call for interest with the Institute of Electrical and Electronics Engineers Inc. (IEEE) 802.3 working group in hopes of getting on the November meeting agenda, says Paul Runcy, the senior product line manager for Broadcom's infrastructure and networking group.

Boiled down, Broadcom's approach takes commercial EPON MAC chips and front-ends them with a prototype PHY layer. If all comes to fruition, the combo could evolve into an ASIC or get incorporated into Broadcom's EPON platform. "That PHY layer could become a product, with a lot of development work," says Broadcom Senior Technical Director Ed Boyd.

Applications for the technology aren't crystal clear, but Broadcom believes what it's cooking up might help cable create a much larger IP pipe that can support multiple services that aren't invented yet. "One target might be to support a Gigabit upstream and multi-Gigabit in the downstream over coax," Runcy says.

Broadcom's not yet ready to say if its technology, which it's showing off in private tradeshow demos, would trump Docsis 3.0, a spec that uses channel bonding to reach shared speeds in excess of 100Mbit/s. Comcast Corp.'s field test at The Cable Show in Chicago demonstrated a 1Gbit/s downstream, and some lab tests have achieved even faster burst speeds. (See Comcast Also Thinking Big With 1-Gig, Comcast Plugs In Cisco for 1-Gig Docsis Demo , Cable Ponders Life After Docsis, The End of Docsis and Speed Thrills .)

"I don't think we've seen the end of innovation on Docsis," Runcy says. For example, cable's already figured a way to bring Docsis-style provisioning to EPON deployments for business-class services.

Still, the work on EPoC is certain to grab the attention of cable engineers who are perpetually seeking ways to avoid taking the expensive FTTP plunge. (See CableLabs Eyes a Super-Sized Upstream and Does HFC Have Plenty of Legs Left? )

— Jeff Baumgartner, Site Editor, Light Reading Cable