Why DOCSIS May Never Die

If you're looking for cable operators to ditch their current hybrid-fiber coax (HFC) networks in favor of FTTH networks en masse, then you'd better prepare yourself for an awfully long wait.

Despite bold, high-profile moves by such major MSOs as Altice to go all-fiber in the US over the next few years, most large US and European cable operators are now lining up in support of the new DOCSIS 3.1 spec to deliver gigabit services. In recent weeks, for instance, both Charter Communications Inc. and Liberty Global Inc. (Nasdaq: LBTY) have signaled their intentions to go ahead with DOCSIS 3.1 rollouts, with Charter issuing an RFP to vendors for D3.1 modems and Liberty Global announcing plans to start D3.1 field trials in the second half of the year. (See Charter Issues RFP for D3.1 Modems and Altice USA Sticks to High-Fiber Diet.)

Beyond that, leading cable operators are also lining up behind the more futuristic Full Duplex DOCSIS spec, which is still bring crafted by CableLabs . Even though that proposed next-gen spec can't possibly be available for field use for at least two more years, such large MSOs as Comcast Corp. (Nasdaq: CMCSA, CMCSK), Charter and Cox Communications Inc. are all counting on it to enable multi-gigabit symmetrical speeds of up to 10 Gbit/s over their HFC networks in the near future.

And even beyond that, cable technologists are now studying how to use coherent optics technology to boost the capacity of their HFC networks much, much further. At their Winter Conference in Orlando last month, CableLabs officials said coherent optics could help cablecos make much more efficient use of the fiber part and optical spectrum of their plant, potentially increasing the total capacity of their HFC networks by more than 1,000 times. (See New CableLabs Optical Tech Promises Big Bandwidth Boost.)

So, exactly two decades after the original DOCSIS 1.0 spec was released by CableLabs, cable's workhorse broadband protocol is still going quite strong today and can look forward to a seemingly limitless future. Happy Birthday, DOCSIS. Welcome to your 20s.

To mark the occasion, we'll be looking closely at the future of DOCSIS when we convene Light Reading's tenth annual Cable Next-Gen Technologies & Strategies conference in Denver next week. The conference, which started out in March 2007 to chart the progress of the then-new DOCSIS 3.0 spec, has now become the premier independent cable tech event in North America. Due to popular demand and a sellout crowd last spring, the event is expanding to two days (Tuesday March 21 and Wednesday March 22) and a larger facility (Curtis Hotel) this time around.

But, unlike that initial conference, we won't be confining our discussions just to DOCSIS' promising prospects. Over the course of the two days, a dozen panels will delve into such meaty topics as cable's pursuit of Distributed Access Architecture, the virtualization of key network functions for both residential and commercial services, and the pros and cons of coherent optics. In addition, we'll tackle cable's deep fiber builds, budding wireless ambitions, IP/cloud video efforts, OTT video maneuvers and 4K/UHD video challenges, along with Full Duplex DOCSIS, DOCSIS 3.1 and other subjects.

Want to learn more about cable's next-gen plans? Sign up now for Light Reading's Cable Next-Gen Technologies & Strategies event on March 21-22, at the Curtis Hotel in
downtown Denver.

Besides all the panel discussions and debates, Cable Next-Gen will also serve up a potpourri of keynote presentations and fireside chats featuring ten tech leaders from the cable, OTT and online video worlds.

Our first-day highlights will include up-close sessions with: Jeff Finkelstein, executive director of advanced access architectures for Cox; Robert Howald, VP of network architecture for Comcast; Dennis Stevens, SVP of product development and strategic programs for Shaw Communications Inc. ; and Balan Nair, CTO of Liberty Global.

And our second-day highlights will include presentations by or one-on-one chats with: Ben Watkins, an executive producer at Amazon Prime Video; Joe Inzerillo, EVP and CTO of MLBAM ; Andrew Ferrone, VP of Pay TV at Roku Inc. ; Jason Thibeault, executive director of the Streaming Video Alliance ; Mike LaJoie, former CTO of Time Warner Cable; and Michael Willner, president and CEO of Penthera Partners Inc. .

Cable Next-Gen will also offer specially targeted breakfast workshops each day. Besides the usual breakfast spreads in the hallways outside the conference rooms, we will have a panel on UHD video's prospects (for service providers only) sponsored by Verimatrix Inc. on Tuesday and a session on why mentoring matters hosted by our Women in Comms group on Wednesday. Both breakfasts require separate registration. (See Mentoring in the Mountains: Join WiC in Denver!.)

As well, the conference will offer lots of schmoozing opportunities for service providers and vendors alike. Besides all the usual breakfasts, lunches and coffee breaks scheduled over the two days, there will be a cocktail reception at the end of the first day, followed by a two-hour party/kickoff event at the nearby Rock Bottom Brewery.

Finally, the weather promises to cooperate with sunny skies and temps in the low-to-mid 60s in Denver next week. So how can you go wrong? Sign up now for Light Reading's Cable Next-Gen Technologies & Strategies event on March 21-22, at the Curtis Hotel in downtown Denver. We'll be expecting you.

— Alan Breznick, Cable/Video Practice Leader, Light Reading

Page 1 / 2   >   >>
Magnets-are-cool 5/31/2017 | 4:39:55 AM
More demanding than one-way-video? I expect that there will be new content and interactivity when facebook sorts out their VR and AR stories. If each VR-user needs 4k video in the down direction and there are 4 users in a household. That's 1Gb to the home used up. If there is AR, then perhaps there is some asstymetrical amount of traffic heading in the other direction as well. But unlikely to be the full 4k up.
kq4ym 3/27/2017 | 4:44:35 PM
Re: "Never" is a very long time It may be hard to predict if and when upload broadcasting by individuals and businesses might be more common than durrent streaming video watching but I would guess it could happen. In that case the fact that "Charter and Cox Communications Inc. are all counting on it to enable multi-gigabit symmetrical speeds of up to 10 Gbit/s," as noted might be a clue that they see the need for those higher speed uploads at least the same speed as downloads as a necessity. 
brooks7 3/17/2017 | 2:06:56 AM
Re: "Never" is a very long time  

Suppose we go full 1984 and stream HD video from every nook and cranny of all space.  Call it 1 HD video per 10 sq ft of the entire globe.

I generally argue that there are not the applications to support business scale bandwidths today or for some time.  But in the long term, there probably will be.  My example is stupid, but you are arguing for 640K of RAM.



Edit:  I thought I would add something about Multi-cast.  That only works with broadcast streams.  Video on Demand, Netflix, and Youtube can not efficiently use multi-cast.   Even subscribers that are watching the same content will be at separate times in the stream.  The data that each sub is receiving is close to unique.
mhui0 3/17/2017 | 1:18:05 AM
Re: "Never" is a very long time I thought streaming is handled by multicast and geographically dispersed storage, at least for people watching video is concerned.


I really doubt bandwidth usage for individuals broadcasting their video will rise to anywhere close to the bandwidth usage of people watching video.
Duh! 3/16/2017 | 5:16:54 PM
Re: "Never" is a very long time That has been my hypothesis for a long time. Growth is sigmoid, not exponential.
brooks7 3/16/2017 | 10:11:29 AM
Re: "Never" is a very long time  


Desktops and Laptops can't really absorb 1 Gb/s and do anything with it effectively.  There is work to do to build applications and cost effective electronics that actually consume even 100 Mb/s on an ongoing basis - not a burst basis.  Today, HD streaming is the biggest bandwidth eater and it is not close to those limits.

The thing is the bottleneck - from a bandwidth standpoint - is not in the access.  And that is the rub.  What we charge people for is the access bit rate.  The bottlenecks are much deeper in the network, where data is more aggregated.  The network was originally built to handle short bursts of traffic with low active periods.  Streaming breaks that latter bit and causes the average bits per second consumed to go up dramatically.  Streaming and even IoT will put pressure on this last bit.

So, do we need more bps at the access?  Not probably beyond 1Gbps for a LONG time.  What does need to happen is that this bandwidth has less oversubscription over time, dramatically less oversubscription...and with more symmetry to the service.

mhui0 3/16/2017 | 2:19:36 AM
Re: "Never" is a very long time Doesn't the ever increasing bandwidth requirements have natural limits too?
brooks7 3/16/2017 | 12:10:42 AM
Re: "Never" is a very long time I think the question we should be asking is whether coax is replaced by fiber or some new wireless.

And I think Coax has a lot longer life yet.  If we go to skinny broadcast bundles backed by switched digital IP video for everything else, then maybe we take 1/2 the spectrum and use it for data.


alanbreznick 3/15/2017 | 5:44:15 PM
Re: "Never" is a very long time I agree with what Dennis said. I also used "never" in the headline because I liked the James Bondian sound of it. Of course, coax and HFC networks will go away someday. But probably not for a long, long time. In the meantime, as stated below, most of us will likely die first. 
mendyk 3/15/2017 | 5:31:32 PM
Re: "Never" is a very long time We're leaving doublespeak behind. It's time for quadruplespeak. Fun times.
Page 1 / 2   >   >>
Sign In