1-Gig Cable Gateway Gets Ready for Its Close-Up

A advanced breed of Docsis 3.0 modem that can get cable within shouting distance of a 1Gbit/s downstream could meet its biggest test yet in November.

Hitron Technologies Inc. plans to submit its gateway to CableLabs for certification testing in wave 96, says Todd Babic, the company's chief sales and marketing officer. According to CableLabs, products for wave 96 are due on Nov. 8, with results expected by late February 2013.

The gateway is based on the Intel Corp. (Nasdaq: INTC) Puma 6 chip that can bond 24 downstream channels and eight upstream channels -- enough to support respective speed bursts of 960 Mbit/s and 320 Mbit/s. The Puma 6 also supports a 16x4 configuration. (See Intel's New Docsis 3.0 Chip Guns for 1-Gig .)

Looking further ahead, Hitron is considering submitting a similar Docsis 3.0 gateway with additional PacketCable VoIP capabilities in wave 98, which is slated to get under way in late February or early March. Hitron's initial batch of gateways based on the Intel chip will support 802.11n, following with faster 802.11ac implementations early next year.

Hitron introduced a standalone 24x8 D3 modem at The Cable Show in May and has more than 100 of them out for evaluation by "most major operators," Babic says. But Hitron opted not to certify that product because its strategy is more focused on Docsis 3.0 gateway devices.

While obtaining certification would give Hitron the OK to sell its new gateway product at retail, the vendor's strategy is to sell directly to MSOs. Even without the retail angle, certification remains an important interoperability hurdle vendors must pass before cable operators will consider products for further testing and eventual purchases. Boiled down, certification "is a license to hunt," Babic says.

But it's unlikely that many cable operators have the channels available or the desire to offer a broadband tier that would fully load a 24x8 modem right away. Hitron believes operators will go for consistency instead, adding channels to the bonding group to stabilize aggregate speeds going for existing tiers, rather than shooting for 1Gbit/s bursts.

Several U.S. MSOs offer D3 tiers that advertise downstream speeds in the range of 50Mbit/s to 100Mbit/s. The extreme case is Comcast Corp. (Nasdaq: CMCSA, CMCSK), which introduced a new tier that maxes out at 305Mbit/s downstream and 65Mbit/s upstream. (See Comcast Revs Up Pricey 305-Meg Tier.)

And deployments of 24x8 modems from Hitron and others will hinge on the readiness of cable modem termination systems (CMTSs), which will need software upgrades to accommodate the latest channel-bonding configurations. Hitron has been demonstrating its new modems and gateways on CMTSs from Casa Systems Inc.

All major CMTS vendors say they are ready now or will be by the middle of 2013 to support broadband modems gateways outfitted with Intel's new chips. Details about their plans to support the latest generation of D3 customer premises equipment (CPE) will be covered in a future story.

— Jeff Baumgartner, Site Editor, Light Reading Cable

Jeff Baumgartner 12/5/2012 | 5:20:31 PM
re: 1-Gig Cable Gateway Gets Ready for Its Close-Up

Got my numbers mixed up. Hitron is targeting certification wave 96 for its first submission of the D3 gateway, not 93, which got underway in July. The numbers should be correct now. JB

Jeff Baumgartner 12/5/2012 | 5:20:31 PM
re: 1-Gig Cable Gateway Gets Ready for Its Close-Up

The other potential timing issue with Hitron's plans has to do with whether CableLabs will even be ready to test modems in this new configuration since the CMTS software for this is still relatively new or still in development, depending on the vendor. 

CableLabs has typically said that it will be ready to test new products when the vendors are ready to have it tested, and it declined to provide any specific comments about Hitron's ambitions for the November cert wave. "As for readiness, we will test to the spec, just as we have done since the inception of certification testing," CableLabs said, but didn't say how 24x8 products might affects its D3 test plans.  We'll know for sure when November rolls around.

And we'll also know by then who else has a 24/8 product on the roadmap... i expect we'll hear about some new models from other vendors at next month's Cable-Tec Expo.  But , at the moment, Hitron appears to be well downstream with its 1st batch of products.



Peyton Maynard-Koran 12/5/2012 | 5:20:29 PM
re: 1-Gig Cable Gateway Gets Ready for Its Close-Up

So if i do my math right, 960 Mbps / 24 Channels = exactly 40 Mbps per channel.  Don't most plants operate at 38.8 Mbps per channel on the downstream?  Wouldn't 24 channels be more like 931.2 Mbps?  And on the upstream 320 Mbps / 8 Channels = 40 Mbps.  Again, don't most upstream plants operate at 30.72 Mbps?  Wouldn't 8 channels be more like 245.76 Mpbs?  And these numbers come from perfect plants. Performance in the field would be roughly 90% of these stated goals.

So are we saying that an MSO would dedicate 24 Channels down and 8 up so that one customer on a particular CMTS port might be able to get 838.08 Mbps down and 221.184 Mbps up? 

Seems really inefficient.

Jeff Baumgartner 12/5/2012 | 5:20:29 PM
re: 1-Gig Cable Gateway Gets Ready for Its Close-Up

Yes, you're right..i was rounding up to 40Mbit/s, so 38.8Mbit/s would get it to 931.2Mbit/s on the downstream.  But that number is also a theoretical high-end burst figure, and not something good expect to see from an advertised speed in a broadband tier. Still, good to clarify the theoreticals and what to expect from a real-world deployment. JB

comtech3 12/5/2012 | 5:20:27 PM
re: 1-Gig Cable Gateway Gets Ready for Its Close-Up

This is a no-sell and wasted product in my opinion because I don't see any MSOs who are worth their salt jumping off a cliff to buy this D3 modem, which would entail giving up a good chunk of their spectrum to support Gbps throughput that the average Joe the plummer don't need! As you and i know, 5-10 Mbps download,and 1-2 Mbps for the average Joe and Jane is more than adequate, and faster speeds are the domain of medical facilities and large to medium size businesses that sends and receives large files,and some of these businesses are already hooked up to dedicated fiber from the MSO's business class services.

monarth711 12/5/2012 | 5:20:21 PM
re: 1-Gig Cable Gateway Gets Ready for Its Close-Up

It is always interesting to see people disregard new products or solution because everyone "knows" that we have enough storage or bandwidth or whatever. Been happening throughout the history of technology and yet people still say it and then look back in 3-5 years and what was enough then seems so slow or small now. After all everyone knew 56kbps modems gave us enough speed back in the day, or DOCSIS 1.0 or VDSL, or ...... I get why people say this type of stuff but unless we revert to the Stone Age again soon, Joe and Jane referenced in the previous post are going to want and likely need more bandwidth.

Also I think folks reading this may miss one of the main benefits of 24x8 devices.  While a few customers may want or need 200, 300, or even 500Mbps, MSOs deploying it and their customers will also see higher aggregate sustained bandwidth to the homes using these devices, not just peak speeds.  In the end giving higher sustained bandwidth to all customers with the new devices may be more of a killer app than giving 1% 800Mbps. 

Take for example the 10Mbps stated below, that means on an 8x4 device today supplying about 250 real Mbps over those 8 channels, only 25 people can get 10Mbps sustained at a time (not 100% true, oversubscription, usage rates, non-peak usage, all can make these calculations more complex but got to simplify so this post is not a mile long).  Depending on the network quality and where in the world you might be there could be 1-500 people sharing that 250Mbps.  Now 10Mbps may seem like a lot but with people streaming video to 1 or more TVs every night that gets eaten up pretty quick.  10Mpbps is perhaps 3 standard def streams or two near HD streams, throw in some other traffic types and we have now broken the 10Mbps threshold.  Or looked at a different way if there are 100 people sharing the 8 channels then everyone can only sustain 2.5Mbps or perhaps 1 stream of video.  Simplified but hope it gets the point across.  So when an operator uses a 24x8 device to move to 12 channels or 18, or 20, up to 24 they are also adding more bandwidth to deliver sustained throughput to the home.  Those same 25 people over say 850Mpbs on a 24x8 device gives each the ability to sustain 34Mbps or 100 people to sustain 8.5mbps.  While this could be done with 3 bonding groups of 8 channels with 8x4 devices, it would not be as efficient or easy to manage as 24 channels together.

To close with a bit more food for thought based on the last post.  Fiber is likely to play a part, and already is for many operators but a few items will keep D3 at 24x8, or perhaps even future 48x24 relevant.  First coax runs into many multiple times more homes than fiber depending on operator.  The cost to migrate from D3 8x4 to D3 24x8 is trivial compared to the cost of running fiber to the home outside of greenfield areas.  Even with some of the newer fiber technologies, materials, and solutions unless it uses an existing delivery method to the home the cost is high.  Fiber deeper in the network and ultimately to the home is possibly the end game and definitely part of the migration but in the meantime DOCSIS has a lot of life left to give.  Another thought, there are a lot of TV channels out there, several 100 from my operator of which I watch about 10.  So for me and what seems like a growing number of people the benefit of adding more channels (HD or SD) versus foregoing those channels for more bandwidth aka giving the channel space over to the data side of the house seems like it will become a real interesting discussion over the next several years.  This will likely further drive interest towards 24x8 devices.

Jeff Baumgartner 12/5/2012 | 5:20:21 PM
re: 1-Gig Cable Gateway Gets Ready for Its Close-Up

I suppose there's a future-proofing angle to this, if there's not a huge delta between 8x4 products  and the new 16x4s and 24x8s. But i agree that we're not going to see msos clammoring to  fill those up in part for the reasons you stated.  And there's are some interesting fiber vs. D3 decisions being made with respect to very fast residential services that we'll cover in some detail on Thurs. morning. JB

Sign In