Networking over coax: evolutions in both LAN and WAN

Article By : Brian Dipert

Enhancing coaxial cable-based WAN and LAN experiences with gear upgrades.

As regular readers may recall, Comcast is my broadband Internet (WAN) service provider. And regular readers may also recall that MoCA is one of the LAN technologies in current use in my home (containing my office), along with 802.11ac mesh wireless and GbE Cat5e and Cat6 wired (powerline isn’t currently in the mix, at least for network packets, although I do use it to distribute dial tone to my office). Spurred on by recent sale prices, I’ve acquired some gear upgrades that I plan to sooner-or-later use to enhance both my WAN and LAN experiences. In this post, I’ll share my rationale for both purchases, along with some background technical info.

LAN first. Some of you may remember that a couple of years ago, one of my initial MoCA adapters, Actiontec’s ECB2200, died courtesy of a close-proximity lightning EMP-induced zap. Since then I’ve relied on a matched pair of Actiontec ECB2500C successors, with the remaining ECB2200 in reserve as backup in case of another lightning strike (all are MoCA 1.1-compliant; as I noted in that earlier writeup, the differences between the two generations are seemingly only cosmetic). Here’s a photo of (a dusty) one of them, connected to the router in the furnace room:

Prompted by a now-expired $116.67-for-two Amazon promotion, however, I’ve bought two pairs worth of their successor, Actiontec’s ECB6200. Not only is the ECB6200 MoCA 2.0-based, for a theoretical per-channel MAC throughput boost from 175 Mbps to 500 Mbps (in “Turbo” mode, 400 Mbps in normal mode), it also implements the specification’s bonded-channels option for twice the theoretical throughput, to a 1 Gbps peak MAC rate (again in “Turbo” mode; even newer MoCA 2.5 further increases this rate to 2.5 Gbps). MoCA 2.0 also reduces latency to 3.6 msec, along with adding lower power consumption standby and sleep modes.

How does MoCA 2 deliver on its bandwidth claims? The details are somewhat nebulous, at least for (I suspect) non-Alliance member companies-under-NDA … both the general information on MoCA’s own website and that of Actiontec are heavy on the what and light on the how. A bit of Google searching, however, has unearthed a few clues … see, for example, MoCA’s “Installation Best Practices for a Home Network” and a 2013-dated presentation to the San Diego Chapter of the Society of Cable Telecommunications Engineers, both available as PDFs.

Somewhat unsurprisingly, considering that Wi-Fi and powerline networking have also gone down similar evolutionary paths, a notable factor in MoCA 2.0’s bandwidth enhancements is its 2× wider RF channel vs MoCA 1.x; 100 MHz vs 50 MHz. As already mentioned, MoCA 2.0’s bonded option also combines two of these wider channels together (MoCA 2.5’s even higher bandwidth derives from merging together 3, 4, or 5 channels’ worth of data, along with other enhancements). I suspect, too, although I can’t find public documentation to support this, that additional throughput boosts come from the use of more modern, more efficient modulation and error detection-and-correction schemes than were employed with MoCA 1.1.

One nice nuance about MoCA 2.0 is that in a mixed-generation adapter topology, MoCA 2.0-compliant devices will still communicate with each other at peak possible speeds even though they need to down-throttle for communication compatibility with legacy MoCA 1.x units. Conversely, with the HomePlug powerline networking standard (for example), the presence of just one legacy-generation device will force the entire topology to operate at that (slower) legacy protocol level. Also, speaking of powerline networking comparisons and contrasts, I double-checked and confirmed that the Actiontec ECB6200 bonded MoCA 2.0 adapter includes a proper GbE transceiver; long-time readers may recall my scorn at seeing claimed multi-hundred-Mbit HomePlug AV adapters constrained by 10/100 Mbps Ethernet ports.

And what will I be using this performance up-tick for? Nothing, admittedly, at least at the moment. Right now, all I’ve got hanging off the non-router end of the MoCA span are two client devices, a Roku streaming media player and an Xbox 360, both only 10/100 Mbit Ethernet-capable (although, with future upgrades already planned, in-between them and the MoCA adapter is a five-port GbE switch). Although I’ve got coax running upstairs to my office directly above the furnace room (i.e. the networking nexus), I also recently replaced the faulty Ethernet cable previously connecting the two rooms with a proper span of Cat 6, so I’m set. But along the same lines, my wife’s office is currently serviced only by Wi-Fi; although I admittedly loath the idea of revisiting powerline networking yet again, there’s a span of coax running outside along the office’s shared wall with the outdoors that I could fairly easily split and turn into a MoCA spur.

Now for the WAN. Although my Comcast “Performance Pro” broadband service tier originally only promised downstream bandwidth of up to 30 Mbps, periodic upgrades have it now assuring speeds up to 70 Mbps, and as you can see, my particular coax tether is even faster than that:

That speed is perfectly acceptable for pretty much everything broadband-related I need … with one exception, which ironically is primarily focused on upstream versus downstream data rates. My “day job” increasingly involves shooting video (with a HD camcorder, as well as capturing the video output by a laptop using a Hauppauge HD PVR Rocket), which I then need to transfer to my “editing guy” via Dropbox.

Uploading a few dozen GBytes of raw camcorder footage takes a long time at 5 Mbps; even doubling that rate (thereby halving the upload time) would make a big improvement. Unfortunately, Comcast’s highest service tier, 1 Gbps, is currently only available in my area as part of a “triple pack” including not only TV but also VoIP, which I’m not interested in. And since Comcast only prominently promotes the “up to” downstream speeds of its various service tiers, not their upstream equivalents, I’m left to rely on the user postings at Comcast’s support forum along with Reddit, DSL Reports, and elsewhere for information, which are inconclusive.

Right, now I’m still using the Linksys DPC3008 I first activated in late 2012; it’s a DOCSIS 3.0-compliant cable modem with support for eight bonded downstream channels and four bonded upstream channels, translating into theoretical speeds of up to 340 Mbps (down) and 120 Mbps (up). It’s always a good idea to have a spare on hand that you can press into service in a pinch, so I’ve also picked up an on-sale Netgear CM500 for $44.95, a DOCSIS 3.0 modem with 16×4 bonded channel support (I’d donated my previous spares to my new-to-Denver nephew and his girlfriend for use in their apartments). And with future Gigabit tier upgrades in mind, I’ve also picked up an on-sale Netgear Nighthawk CM1100 for $119.99.

The CM1100, a DOCSIS 3.1-compatible modem (backwards-compatible to DOCSIS 3.0 and below, of course), is an interesting design. Netgear calls it a “Multi-Gig” modem, which begs the obvious implementation question: did the company put an advanced 2.5 or 5 GbE transceiver in the product, or (as with previously-mentioned powerline networking) is this just a “smoke and mirrors” spec? Neither, actually; the CM1100 includes dual 1 GbE ports. The practical usefulness of this potential is admittedly limited; only a handful of consumer (or even business, for that matter) routers support dual WAN connections, and then mostly only for fail-over support (cable as primary and DSL as backup, for example), while even fewer support bonding together those dual WAN ports for higher effective WAN throughput. Still, the CM1100 will work just fine as a 1 Gbit cable modem with only one Ethernet port in use; keep in mind that 1 Gbit DOCSIS 3.1-compliant modems usually sell for $200 or more. I suspect that widespread consumer adoption of “Multi-Gig” technology awaits widespread implementation of 2.5 and 5 GbE ports in modems and routers.

My enthusiasm for a broadband cable service tier upgrade is mixed. On the one hand, as previously mentioned, I’d welcome any uptick in upstream bandwidth, though my research so far doesn’t suggest I can guarantee that this will actually occur. On the other hand, I might end up more quickly creeping up toward my monthly bandwidth cap than before, for reasons that I don’t care about at all. Media streamers such as Rokus and Amazon Fire TV devices will by default automatically select from an assortment of available video resolution and bitrate options (at Netflix and elsewhere) based on how much downstream bandwidth is available at your connection. More bandwidth potential = a larger download payload, even if you’re perfectly happy with the image and audio quality you’d had before. Plus, of course, sooner-or-later (i.e. once any promotional-pricing period ends) I’ll also be paying more for broadband than I was before.

One other comment, in closing. As it turns out, the LAN and WAN sections of this post are related. As MoCA’s Wikipedia entry notes:

In March 2017, [the] SCTE/ISBE society and MoCA consortium began creating a new “standards operational practice” (SCTE 235) to provide MoCA 2.0 with Docsis 3.1 interoperability. Interoperability is necessary because both MoCA 2.0 and Docsis 3.1 may operate in the frequency range above 1 GHz. The standard “addresses the need to prevent degradation or failure of signals due to a shared frequency range above 1 GHz.”

And with that, I’ll conclude for today. Comments are as-always welcomed (selfishly, I’d be particularly interested in feedback from those with Comcast service tiers higher than mine on the upstream bandwidth they see!).

Brian Dipert is Editor-in-Chief of the Embedded Vision Alliance, and a Senior Analyst at BDTI and Editor-in-Chief of InsideDSP, the company’s online newsletter.

Related articles:

Leave a comment