Telcos that have been deploying fiber broadband are having a moment in the sun, finally reversing years of broadband subscriber losses to the cable companies. But will this last?
What has given telcos an edge is fiber broadband’s ability to support gigabit and even multi-gigabit speeds bi-directionally. With traditional technology, cable companies can offer gigabit and multi-gigabit speeds but only in the downstream direction. Upstream speeds sometimes top out at just 20 Mbps.
All that is set to change, though. Telecompetitor talked recently with Curtis Knittle, vice president of wired technologies for Cablelabs, about the technology that’s coming soon for cable. Upstream speeds are poised to get a lot faster – maybe not fully symmetrical, but probably fast enough for most consumers.
Perhaps of greater concern, it’s still a lot easier to upgrade cable hybrid fiber coax (HFC) infrastructure than it is to replace telco copper with fiber. (HFC networks use fiber networks from the cable company headend to a node where signals are converted to coaxial cable for delivery to customer locations.)
To maximize speeds, cable companies will be deploying DOCSIS 4.0, the next generation of the DOCSIS standard. They also will be deploying, and in some cases already have deployed, distributed access architecture (DAA). In addition, the cable companies are expanding the amount of spectrum that is available for use within the coaxial portion of the link and making more of that spectrum available for upstream communications.
Here’s what we learned from Cablelabs about all this.
DOCSIS 4.0 and DAA
When we talked to Knittle last month, production silicon supporting the DOCSIS 4.0 standard was expected soon. Cablelabs is getting set to certify devices and Knittle expects to see “wide-scale” deployments of the technology beginning later this year.
Another key upgrade is the use of distributed access architecture (DAA), which as Knittle explained, involves “moving intelligence to the edge” using Remote PHY and Remote MAC PHY devices.
Those devices are installed in the node where optical transmissions are converted to coax. They improve the signal-to-noise ratio (SNR), enabling the cable network operator to boost speeds and capacity.
“Between those devices and the hub, we’ve converted to digital optics in the form of Ethernet – 10 Gig Ethernet or 25 Gig Ethernet,” Knittle said.
Putting it another way, he said, “distributed access architecture is defined by digital optics and RF generation at the edge.”
It’s worth noting that some cable companies have been deploying DAA for several years as it provides benefits even with current generation DOCSIS 3.1 equipment.
High-Splits, Etc.
For years, cable companies relied on no more than 860 MHz of spectrum within the coaxial portion of their HFC networks. Increasing that spectrum can boost speeds.
Many cable companies are opting to expand the useable spectrum to 1.2 GHz or even 1.8 GHz and to expand the portion of the band used for upstream communications.
As they expand spectrum, Charter and some other companies plan to use what Cablelabs calls “high splits.” With this approach, the amount of spectrum available for upstream bandwidth is greater than in the past but is still considerably less than what is available for downstream communications.
For example, Charter says it expects to support speeds up to 5 Gbps downstream using DOCSIS 4.0 modems with 1.2 GHz of spectrum and high splits. The company expects to support speeds up to 10 Gbps downstream using DOCSIS 4.0 with 1.8 GHz of spectrum. In either case, upstream speeds will be 1 Gbps.

It’s worth noting that Charter’s target speeds may not represent the maximum speeds that the technology can support. Instead, they are the speeds that the company sees as appropriate in today’s market.
One cable operator – Comcast—is making a different choice as it expands spectrum. It will be using a full duplex approach.
With a full duplex approach, the total amount of spectrum available within the coaxial portion of the link will be 1.2 GHz. The key difference is that a large portion of the spectrum will be available for either upstream or downstream communications, which means that Comcast will be able to offer higher upstream speeds.

Comcast says it has achieved speeds of 6 Gbps downstream and 4 Gbps upstream using full duplex DOCSIS 4.0 in a real-world test. The company expects to be able to replicate that across a large part of its network without taking fiber deeper into the network beginning later this year.
It’s important to note that the speeds that Comcast has been able to achieve do not necessarily represent the speeds that the company plans to offer. The company has not been forthcoming with details there.
It’s also important to note that, as with DAA, cable companies already have been expanding spectrum overall and on the upstream, and as they do, those moves also provide advantages even with current generation DOCSIS 3.1 equipment.
Fast Upgrades
Charter recently announced plans to upgrade its entire serving area by 2025.
Not everyone will get DOCSIS 4.0. The company plans to do its upgrade in three stages, with different elements of the technologies we’ve discussed implemented at each stage.
As the upgrades are completed, 15% of the company’s customers will be able to get speeds of 2/1 Gbps, 50% will be able to get speeds of 5/1 Gbps and 35% will be able to get speeds of 10/1 Gbps.
Comcast expects to upgrade its entire customer base to DOCSIS 4.0 by 2025.
For telcos, the process of upgrading existing DSL networks to fiber broadband takes longer and, in some cases, the upgrade to fiber may not be cost-effective. Charter is targeting $100 per location for its upgrade plans and, according to financial analysts at SVB MoffettNathanson, Comcast is targeting $200 per location. Upgrading DSL networks to fiber is considerably more costly.
It’s worth noting that when cable companies expand to areas they never served previously, they are deploying fiber broadband. And as companies like Charter and Comcast pursue projects in unserved and underserved rural areas — projects funded, in part, through government programs — we’ll see them use fiber there, too.
We also may see them selectively upgrade some of their existing infrastructure to fiber broadband.
Nevertheless, their current HFC infrastructure may soon be a lot more competitive than it has been in the last year or two. And it could remain that way for quite some time.
Updated to change MoffettNathanson to SVB MoffettNathanson and to state that deploying DAA improves the signal-to-noise ratio
Lets also not forget that companies like comcast are keeping your average Joe (owns their own modem) from utilizing mid-split for upstream, I’d also assume they will do the same when DOCSIS 4.0 come around. They will limit it to their own hardware which comes at an additional monthly cost of $25. This puts their price point well above their fiber competitors. It’s no wonder they have had 2 bad quarters. Not only will they provide you with a product inferior to FTTP, they’ll charge you more for it.
Yes the cable companies are trying to remain relevant. Instead of just biting the bullet and doing as much fttp as possible they keep trying to polish up coax. This time they are screwing up royally because most fiber providers offer at least 1g symmetrical already. The next leap will be between 1g and 10g which is possible on the current infrastructure. It will be end of 2024 that coax even supports 2g/1g anywhere near symmetrical. Then they keep saying “this is what customers need” no this is all they can provide. If they had played their cards right they would have enabled Dosis 3.1 upstream and let that be the end of coax.
Agreed! This piece meal approach by the cablecos will ultimately reach a breaking point because although being scalable, there is a limit to where you can squeeze the juice out of coax before it gives up on you. As a matter of fact, by increasing the spectrum to a gigerherz means adding more inline amps,or line extenders to regenerate the signal due to span loss. Adding more amps increases the noise floor each time one is added to the cascade of line extenders and mini bridge amps. That is a line technician’s headache to find fault,and to balance these amps.The cablecos may be running from the initial cost of going full FTTx,but in the long run it is cheaper with low maintenance given that they don’t,or probably won’t need line techs because there are no inline amps or actives for miles, almost zero reliance of the electric power companies to keep their backbone running. Besides, fiber optic cable and the cost to install it has gone down tremendously in years. FTTx cost to reliability ratio is far superior to what the cablecos are doing because of less down time caused outages that occur in the cable plant, and trust me, they’re many. Even routine maintenance that may involved rebalancing an amp, a tap faceplate, can bring down an entire neighborhood. For FTTx, it may gone down if they is a catastrophic break in the cable cause by a tree falling down or motor vehicle bring the pole down. Now, this is where HCF wins the restoration war if it only involves the coax portion and not the fiber that goes to the node. The break can be temporarily jumped until a permanent restoration is done.
And that is why Cable Companies may close the tech gap but not the customer buy-in or gain fans from the consumers.
Fix your pricing model and you will find very loyal subscribers. Not everyone wants to pay a mortgage payment for internet or basic TV service.
Better question is when will these higher gigabit & multigigabit speeds be AFFORDABLE…
The consumer hasn’t really seen widespread gigabit under $50… let alone multi gigabit under $100.
As someone that has worked in a datacenter. Docsis will never be able to even shed a light on fiber. Currently a single fiber strand can to 19.2 terabit using DWDM.