Light Reading

EPON-Over-Coax Starts Its Standards Journey

Jeff Baumgartner
LR Cable News Analysis
Jeff Baumgartner
1/11/2012
50%
50%

If a new Institute of Electrical and Electronics Engineers Inc. (IEEE) project bears fruit, cable operators might be able to delay costly fiber-to-the-home (FTTH) deployments, or even avoid them altogether.

The proposed physical-layer standard is called EPON-over-Coax (EPoC), and its goal is to provide symmetrical 10Gbit/s speeds over hybrid fiber/coax (HFC) networks. It's just getting started; the first EPoC Study Group meeting will get together on Jan. 24 in sunny Newport Beach, Calif.

Here's the slide deck (PDF) that was presented during an IEEE 802.3 call-for-interest meeting last November. The IEEE voted at that meeting to form the study group, a move that simply "gives us the right to study the problem," says Study Group Chairman Howard Frazier, who also serves as senior technical director at Broadcom Corp. (Nasdaq: BRCM), a chipmaker that has been championing EPoC. (See Broadcom Crafting PON-Speed HFC.)

EPoC will need to be compatible with or work around existing cable services, which tend to vary from operator to operator and from HFC plant to HFC plant.

"There's no way operators will do a forklift upgrade. We'll have to adapt to what they're deploying five years from now," Frazier says. That means EPoC will have to coexist with QAM-based video systems, Docsis 3.0 platforms, cell backhaul deployments and the next-gen Converged Cable Access Platform (CCAP), which some view as a possible bridge to EPoC. (See Comcast Gets Ready for CCAP and Cable Rethinks 'Modular' CCAP .)

Here's an example of how EPoC spectrum might live alongside cable's other services:



QAM's big goodbye?
EPoC could begin to steer cable away from its traditional QAM modulation schemes, and this part of the discussion is expected to be among the most hotly debated as engineers mull ways to modulate Ethernet on coax efficiently.

One idea that will get much attention is orthogonal frequency-division multiplexing (OFDM), a scheme popular in the wireless world that could help cable pump out more bits per hertz than they do today with QAM, says Shane Eleniak, vice president of advanced broadband solutions at CommScope Inc.

CommScope has a chip team (that came on board via the company's purchase of edge QAM maker LiquidxStream Systems Inc.) that's been eyeing the development of a mixed-mode QAM/OFDM Application-Specific Integrated Circuit (ASIC) or a more flexible Field Programmable Gate Array (FPGA) implementation, he notes. (See Why CommScope Bought LiquidxStream.)

Frazier acknowledges that OFDM is one scheme that the working group will likely consider, but he warns that nothing's been decided. "We're going to have a very interesting discussion," he says. "We'll have to duke it out and see what works best."

But the cable industry has definitely fallen in love with EPON. Following heavy adoption of EPON by service providers in Asia, domestic cable has grown partial to the technology in recent years, increasingly relying on it to deliver business services. Taking things a step further, CableLabs has also created a spec that grafts Docsis-style provisioning to EPON services. (See Docsis Gets Its EPON On.)

EPoC timetables
Based on the usual IEEE timetable, an EPoC standard would take about three years to complete. Frazier expects some pre-standard products to start coming out in late 2013 or early 2014. "There will be trials ... but I'm thinking that the deployment horizon will start two to three years out."

He sees initial adoption of EPoC happening in North America and also in China, where it's estimated more than 190 million users are connected to coax today, with 300 million expected to be passed by HFC by 2019, according to China's State Administration of Radio, Film, and Television (SARFT).

Here's who's supporting EPoC so far:

Table 1: EPoC Dudes

Alcatel-Lucent Aurora Networks Inc. Bright House Networks
Broadcom Corp. Cogeco Cable Inc. CableLabs Inc.
Comcast Corp. Cox Communications Inc. Dell Inc.
FiberHome Telecommunication Technologies Co. Ltd. Harmonic Inc. HP
High Speed Design Huawei Technologies Co. Ltd NeoPhotonics Corp.
PMC-Sierra Inc. Qualcomm Inc. Sumitomo Electric Industries
Technical Working Committee of China Radio & TV Association Time Warner Cable Inc. Wuhan Yangtze Optical Technologies Co. Ltd.
ZTE Corp.
Source: IEEE




— Jeff Baumgartner, Site Editor, Light Reading Cable

(8)  | 
Comment  | 
Print  | 
Newest First  |  Oldest First  |  Threaded View        ADD A COMMENT
Duh!
50%
50%
Duh!,
User Rank: Light Sabre
12/5/2012 | 5:46:03 PM
re: EPON-Over-Coax Starts Its Standards Journey


... not that it matters.


Why would an MSO go to a different PMD/MAC on legacy HFC plant?  Expecially since DOCSIS is reasonably well optimized for the HFC environment, and EPON was designed for PON.  So, for example, DOCSIS can use contention/reservation for upstream grants because receiver overload is not damaging.  GPON has to use polling/reservation, which is fine because of the relatively small split ratios in the PON, but would (I suspect) tend to be pretty inefficient with a 125:1 split or bigger.


It looks like this scheme is going to require replacing every node, amplifier, and passive in the network to get above 1GHz.  Not that this is unprecedented, but experience shows that this is a non-trivial and expensive enterprise.   We've seen several schemes like this fail to get traction because of the cost of the upgrade.    And it looks like the upstream/downstream bandwidth asymetry gets worse rather than better.  So if you're going to forklift anyway, why not move the high/low split?


DOCSIS 2.0 and above have a more specrtrally efficient upstream modulation scheme, S/CDMA.  How it might compare to any OFDM scheme that 802.3 might come up with is not clear.   While it's likely that there are more spectrally efficient downstream schemes than bonding together 6MHz channels of 64-QAM, I'm not quite seeing downstream efficiency as the long pole in the tent.


It's also not unprecedented for the industry to get distracted by shiny new objects.  Or start up new standards activities with no serious plan to deploy them.


Perhaps somebody with better insight can explain?

Carol Wilson
50%
50%
Carol Wilson,
User Rank: Blogger
12/5/2012 | 5:45:58 PM
re: EPON-Over-Coax Starts Its Standards Journey


I'm not a cable guy and I can't address the issues raised in terms of the modulation schemes or the cost of the upgrades, but in the telecom world, the biggest chunk of change in deploying PON was always  the cost of putting in fiber optic cable to every home. Wouldn't those economics play out here as well?

comtech3
50%
50%
comtech3,
User Rank: Light Beer
12/5/2012 | 5:45:56 PM
re: EPON-Over-Coax Starts Its Standards Journey


In Chestnut Hill,Pennsylvania, there are literally no LEs (line extenders).It is all optical nodes in that Comcast service area because that's why Ralph Roberts and other big wigs in the company live,and use to live.Anyway, my point is that cablecos can do away with LEs all together and allow a single node  to serve about 250 subs.The taps at each location can be converted to optical transceivers and be powered by the same existing hardline coax from the node.In other words, the electronics inside the housing of the node would contain a fiber port and a power output to power the optical transceiver taps.The taps would convert light to RF and vice versa because they would still use the existing coax drop to the home.However,for plant extension, a reserve optical port would be provided.


The cost involved may not be as labor intensive as Verizon Fios because the fiber is already there at the node.The only problem is getting this done where existing plant is underground,and permission for digging up properties and roadways may result in several stumbling blocks as Verizon has found out the hard way.


Despite the affore mentioned diffulties, the approach may be beneficial in the long given the there would be less power consumption from cascading amplifiers, less noise as there won't be any other noise figure (NF) than what is produced by the node.In terms of power outages, the maintenance technician would have a single amp to contend with to apply standby powering.This approach may even be better than RFoG( RF over glass.) Also with no amp cascades, the cablecos will not only able to exceed a 1GHz,but able utilize almost all of the sub-band frequencies.


What do you think?

Halichopter
50%
50%
Halichopter,
User Rank: Light Beer
12/5/2012 | 5:45:46 PM
re: EPON-Over-Coax Starts Its Standards Journey


I think the appeal of EPoC will vary depending on the plant and the channel line up at the particular MSO/region and so will the optimal spectrum splitting. The Chestnut Hill scenario would seem ideal for an upgrade like this. The problems that the initial post raised from the interestingly named "Duh!" are real. However they apply as well to any solution that requires an increase in the upstream spectrum, whether it be more DOCSIS 3.0 channel bonding or a new MAC/PHY like EPoC.  The fewer actives between the node and the subscriber the easier it will be to make the change. The main problem with the current HFC plant spectrum allocation is that the upstream bandwidth is very small compared with the downstream and it is the lowest SNR spectrum available.  In some cases the 5MHz to 12MHz bandwidth is not really useable or only usable with low bits/Hz modulation such as QPSK. This leaves only 42MHz-12mMHz = 30MHz of bandwidth vs > 700MHz for downstream. It is changing the amplfiers, lasers, and diplexers to fix this problem that costs money.


The question of whether the EPoC spectrum must be placed above 1GHz depends on whether the additional upstream bandwidth comes by expanding the current upstream band vs placing it above 1GHz or at high frequencies and also whether bandwidth below 1GHz can be reclaimed via elimination of the analog channels or replacement of the QAM channels. Again, probably the answer differs from one MSO to another.


At first like "Duh!" I was puzzled by the proposal for yet another new modulation scheme beyond the well established DOCSIS 3.0. Do we really need another way of sending data over coax? It is true that some more bandwidth can be squeezed from the plant by using OFDM. There are proposals to use OFDM QAM constellations as large as 4096. However this requires a signal to noise 12dB higher than needed to support the current QAM-256 and only results in 50% more bandwidth (apples to apples not counting elimination of wastful D3.0 guardbands etc.)


I think that the more important consideration may be the cost of the electronics at the headend and home. PON equipment has already been reduced in cost to consumer levels due especially to price pressure from China operators. Considering an EPoC media converter would be shared by multiple homes, the already low PON electronics/optics cost/sub will be further divided.  OFDM is now low in cost since it became the universal wireless solution for WiFi and 4G. Putting the two together could result in a low cost solution compared to DOCSIS 3.0.


Of course everything depends on volume. If the EPoC solutions become splintered by using many different spectral plans or worse, multiple PHYs, and if EPoC is only embraced in the US and not world-wide then the cost reduction may not be compelling.


Time will tell.

Halichopter
50%
50%
Halichopter,
User Rank: Light Beer
12/5/2012 | 5:45:45 PM
re: EPON-Over-Coax Starts Its Standards Journey


We seem to be on the same wavelength (or should I say frequency?), "Duh!"

Duh!
50%
50%
Duh!,
User Rank: Light Sabre
12/5/2012 | 5:45:45 PM
re: EPON-Over-Coax Starts Its Standards Journey


Halichopter,


I was thinking that the only plausible rationale might be something along the lines you're talking about:  effectively, EPON with coax drops. 


I suppose something like that might be feasible.  Trying to overlay it on existing services will be difficult and expensive.   Building a very low cost, energy efficient media converter + optical power splitter that will be a drop-in replacement for (and act as) a multi-tap will be a challenge.  I image that such a device will have somewhat more RF insertion loss than a multi-tap, which may or may not be a problem.   Plus there will be challenges with powering and ODN engineering for the distributed split.  And big construction expenses to overlash fiber on coax.


If that is the architecture they have in mind, it will be interesting to see it play out.


 

Duh!
50%
50%
Duh!,
User Rank: Light Sabre
12/5/2012 | 5:45:45 PM
re: EPON-Over-Coax Starts Its Standards Journey


That wasn't the question.  Without a doubt,  an incumbent with existing coax plant would hope to defer investment in brownfields FTTH infrastructure for as long as possible.  And bandwidth(Hz)/reach curves for coax puts them in a favorable position for doing so, relative to incumbents with existing copper pairs.  


There are lots of good ways for MSOs to increase the downstream bandwidth (and SNR) available for DOCSIS-based services.   They are doing most of them: eliminating analog channels, plant clean-up, SDV, 1GHz upgrades, node segmentation, "fiber rich" architectures.   The long pole in the tent is the limited amount of upstream bandwidth, which is a much more difficult problem.


What I'm seeing here doesn't seem to address those problems.   In fact, I'm not clear as to what problems it does address.


I'll be interested to see Jeff's update.

From The Founder
Light Reading sits down at CES with the head of Cisco's service provider video business, Conrad Clemson, to discuss how NFV and cloud security relate to video, the challenge of managing 4K/8K traffic, the global expansion of Netflix and virtual reality.
Flash Poll
Live Streaming Video
CLOUD / MANAGED SERVICES: Prepping Ethernet for the Cloud
Moderator: Ray LeMaistre Panelists: Jeremy Bye, Leonard Sheahan
LRTV Custom TV
Join Us at the Digital Operations Transformation Summit

2|4|16   |   03:52   |   (0) comments


The Digital Operations Transformation Summit on February 21, 2016 at the Crowne Plaza Barcelona Fira Centre will bring together 50 senior executives to engage in a unique debate on the opportunities and challenges presented by the transformative evolving digital landscape. RSVP now at events@lightreading.com.
LRTV Custom TV
Making the Test: ADVA Ensemble Connector vs. Open vSwitch

2|4|16   |   01:28   |   (0) comments


Light Reading, in partnership with EANTC, recently tested ADVA's Ensemble Connector, which replaces open vSwitch and offers carrier-grade capability and interoperability. The test results strengthen ADVA's credibility as a provider in the virtualization space.
LRTV Custom TV
Bridging the Gap Between PoCs & Deployment in NFV

2|4|16   |   31:50   |   (0) comments


Charlie Ashton of Wind River presents the keynote at Light Reading's 2020 Vision executive summit in Dublin.
Between the CEOs
CEO Chat With Mike Aquino

2|3|16   |   17:34   |   (0) comments


The former CEO of Overture Networks, Mike Aquino, discusses why truly open virtualization solutions provide service providers with the greatest choice.
Shades of Ray
MWC: Buckle Up for 5G & the IIoT

2|2|16   |   02:28   |   (0) comments


This year's Mobile World Congress looks set to be a 5G land grab and a chance to get down and dirty with the Industrial Internet of Things (IIoT) – but what will the 5G discussions actually be about?
LRTV Custom TV
Case Study: Building China's Next-Gen TV Networks

2|2|16   |   5:01   |   (0) comments


With over 2 billion viewers worldwide, Shenzhen Media Group is one of China's largest content producers. By partnering with Huawei and Sobey, SZMG was able to modernize media operations with the Converged News Center, a production studio that is a model for next-generation workflows.
LRTV Custom TV
Quad Channel Modulator Driver with 46 Gbaud Capability from MACOM

1|28|16   |     |   (0) comments


MACOM's MAOM-003427 is the industry's first surface-mount modulator driver with 46 Gbaud capability to support next generation 200G and 400G applications.
LRTV Custom TV
Video Infographic: Validating Cisco's NFV Infrastructure

1|26|16   |   02:24   |   (1) comment


We all know that the network of the future will be virtual, but when will virtual become a reality? This video infographic covers the four key areas in which Light Reading, in partnership with EANTC, tested Cisco's NFV infrastructure: performance, reliability, multi-service capabilities and single pane of glass management.

For the full report, see

Between the CEOs
CEO Chat With Phil McKinney, CableLabs

1|22|16   |   13:36   |   (1) comment


At CES in Las Vegas, we met with Phil McKinney, CEO of CableLabs. Phil provides an update on the rollout of DOCSIS 3.1, his views on the future of open source and how consumer interest in virtual reality could affect network traffic.
Between the CEOs
Ericsson CTO on the Changing Telecom Market

1|21|16   |   10:26   |   (0) comments


At CES 2015, CTO of Ericsson, Ulf Ewaldsson, sits down with CEO of Light Reading, Steve Saunders, to discuss the changing telecom market, the new partnership with Cisco and the future of the telecom industry.
LRTV Interviews
Ireland's Data Dream

1|21|16   |   14:31   |   (0) comments


Host In Ireland president Gary Connolly tells Light Reading's Steve Saunders about the role Ireland is playing in hosting data for the world's largest organizations.
LRTV Custom TV
Brocade Keynote: Transitioning to the New IP

1|20|16   |   27:23   |   (0) comments


At 2020 Vision in Dublin, Andrew Coward, VP of Service Provider Strategy at Brocade, presents the transition to the New IP.
Infographics
Cisco's latest VNI numbers suggest the world will be using 366.8 exabytes of data on smartphones and Internet of Things devices, up from 44.2 exabytes, in 2015.
Hot Topics
Alphabet Is Serious About Google Fiber
Mari Silbey, Senior Editor, Cable/Video, 2/1/2016
Did Juniper Pay 'Peanuts' for BTI?
Mitch Wagner, West Coast Bureau Chief, Light Reading, 2/2/2016
Google's 5G Radio Ambitions Are Expanding
Dan Jones, Mobile Editor, 2/5/2016
How Data Center Outsourcing Fuels AT&T NetBond Growth
Carol Wilson, Editor-at-large, 2/3/2016
3.5GHz Startup Gets $22M for Small Cells
Dan Jones, Mobile Editor, 2/2/2016
Like Us on Facebook
Twitter Feed
Webinar Archive
BETWEEN THE CEOs - Executive Interviews
The former CEO of Overture Networks, Mike Aquino, discusses why truly open virtualization solutions provide service providers with the greatest choice.
As anyone who knows me will tell you, I like to think I know a fair bit about this next-gen-comms malarkey, but there's nothing like an interview with one of the ...
Animals with Phones
Happy Groundhogs for Technology Day! Click Here
Live Digital Audio

Broadband speeds are ramping up across Europe as the continent, at its own pace, follows North America towards a gigabit society. But there are many steps to take on the road to gigabit broadband availability and a number of technology options that can meet the various requirements of Europe’s high-speed fixed broadband network operators. During this radio show we will look at some of the catalysts for broadband network investments and examine the menu of technology options on offer, including vectoring and G.fast for copper plant evolution and the various deployment possibilities for FTTH/B.