SCTE/ISBE CABLE-TEC EXPO – DOCSIS 4.0, distributed access architectures and virtualization all promise to beef up cable network capacity, enhance security, deliver low-latency capabilities and improve network efficiency, but they also will lead cable operators to mobile and, ultimately, to service and network convergence.
That's according to a reading of the technology tea leaves by John Chapman, a Cisco fellow and the company's chief technology officer for cable access.
"Today's cable operators are becoming mobile operators," he declared this week in an exploratory workshop focused on "what's next" for the cable industry.
Chapman also points out that plenty of cable operators around the globe are already in the mobile business – as outright MNOs or, in the case of Charter Communications, Comcast and Altice USA, in the mobile game through MVNO agreements, with the potential to be so much more.
As for that "much more" bit, Chapman also notes that US cable operators played a major role in the recent bidding for 3.5GHz CBRS spectrum. Three of the top five bidders (Charter, Comcast and Cox Communications) were cable ops, and seven of the top 25 were cable operators.
Wireline begets wireless
Gaining access to licensed CBRS spectrum puts certain cable operators in position to be mobile network operators as well, but it will be their respective wireline networks that will underpin that convergence strategy, Chapman said.
Some of that potential ties into cable's pivot toward a distributed access architecture. "It's a way of rebuilding the plant," Chapman said, noting that a big piece of it involves a shift from proprietary analog fiber to digital and IP-over-Ethernet.
"You get to bring Ethernet to the neighborhood and DOCSIS to the door," he said. Converting the HFC network to an IP network, he adds, is a "very powerful construct" because operators can use it not just to drive DOCSIS, but also to drive PON, small cell deployments and fixed wireless access.
That, Chapman predicts, will lead to subscriber convergence and, later, service and architecture convergence. On that latter piece, Chapman put forth the notion that the cable modem termination system (CMTS) could be "rebuilt" to be part of a mobile system that uses the 5G core to manage data. But whether that makes everything more complex or less is still up for debate, he said.
But HFC – thanks in part to its access to power, its backhaul capabilities and widespread location advantages – is well equipped to support the massive number of 5G small cells that will be required, Chapman said. He held that it could take about 125 to 200 CBRS small cells to replace a microcell, and as many as 175,000 millimeter-wave small cells.
But the economics have to work. In the CBRS case, the cost of deploying a small cell would have to be 1% of what they are in a microcell example. There's some good news there for MSOs – if a CBRS small cell installation needs to cost between $100 to $250, that's about the same cost of a cable modem install, Chapman reckons.
"It cannot be business as usual," he said of the financial challenge faced by the need to deploy gobs of small cells. "It's not just a matter of putting up 200 towers instead of one tower."
Plus, a 100Mbit/s CBRS small cell roughly equates to the data supported by an entry-level cable modem. "So there's a very nice match between the speed of a small cell and the speed of a cable modem," he said. "The bandwidths match and the economics look very interesting" for a deployment of a small cell in the home or in a strand-mounted device hanging off a distributed DOCSIS network.
HFC also has some cards to play with the timing and low-latency capabilities that mobile networks require. Enhancements to DOCSIS, including the DOCSIS Time Protocol and the CableLabs SYNC specification, are there to help fit the bill, Chapman said.
"It turns out, behind every wireless network is a great wireline network," he said. "And the HFC plant with DOCSIS is a really great wireline network."
Hitting the DOCSIS sweet spot
Chapman's talk also explored the various options and flavors available to cable operators for where to go next with DOCSIS 3.1 and DOCSIS 4.0.
DOCSIS 3.1 provides operators with a way to extend spectrum to 1.2GHz and pair it with a beefier swath of upstream spectrum – a "mid-split" to 85MHz or a "high split" up to 204MHz.
DOCSIS 4.0 has even more optionality – the use of Full Duplex DOCSIS (which raises the spectrum ceiling to 1.2GHz with the ability to run upstream and downstream traffic in the same block of spectrum), or Extended Spectrum DOCSIS (lifting the spectrum ceiling to 1.8GHz, but keeping the downstream and upstream traffic separated, with the ability to raise the upstream as high as 684MHz).
Chapman notes that operators have ten choices with respect to frequency changes, so what's the best choice?
Much factors into that question and will vary by operator. But, based on Chapman's analysis, the "sweet spot" for DOCSIS 3.1 is plant built out to 1.2GHz with a high-split upstream – offering 8.6 Gbit/s of downstream capacity as well as 1.4 Gbit/s of upstream (or 2.8 Gbit/s with two upstream ports). This best-case scenario assumes that the operator gets rid of its legacy QAM video and goes with all-DOCSIS and IP delivery for video. That set-up would enable an operator to use DOCSIS 3.1 at its "maximum designed bandwidth," Chapman said.
The DOCSIS 3.1 "runner-up" is the 85MHz upstream use case. That will give the operator the ability to deliver upstream capacity in the hundreds of Mbit/s, but not 1-Gig.
"The 204 [MHz high-split] option is really where the money and the opportunity is for DOCSIS 3.1," Chapman said.
The Cisco fellow views the sweet spot for DOCSIS 4.0 as plant built to 1.8GHz with an upstream set at 396MHz, providing 12 Gbit/s of downstream capacity and 3 Gbit/s of upstream capacity (or 6 Gbit/s with the aforementioned two-port upstream).
The runner-up for DOCSIS 4.0 is sticking with a 204MHz return, he said, noting that this option "shines" if an operator still has QAM video running on the plant.
Long live HFC
Although some cable ops, such as Altice USA, are favoring FTTP overbuilds/upgrades, Chapman believes that there's still plenty of capacity to mine from HFC.
"We have a 100x growth path in the cable plant," he said.
As he explains it, HFC can go up to 14 Gbit/s with DOCSIS 4.0, roughly a 10x growth path. Operators can also continue to segment their plant with node splits and put fewer and fewer customers on a node – down into the region of about 50 homes sharing a node, representing a factor of ten if 500 homes per node is the starting point. Combined, that's how he gets to 100x.
"The beauty … is that the bandwidth is there, but you don't need to turn it on until there is a business case. So it's pay as you grow."
- Verizon, Dish, Charter, Comcast: Here are the 20 biggest CBRS auction winners
- CableLabs starts to connect on network convergence
- Cable's 'GAP' initiative feeds urge-to-converge network
- 'Shaw Mobile' banks on network convergence
- More spectrum the key to DOCSIS 4.0
- CableLabs reaches key milestone on path to '10G'
— Jeff Baumgartner, Senior Editor, Light Reading