Is it time to upgrade to mesh networking?

-October 23, 2017

In case you haven't already also noticed, right now's a great time to buy a new router. Here are just a few recent examples that have come across my various RSS feeds:
  • A TP-Link Archer C5 (AC1200) for $49.99
  • A (highly rated) TP-Link Archer C7 (AC1750) for $79.96
  • A TP-Link Archer C9 (AC1900) for $84.99
  • A TRENDnet TEW-818DRU (AC1900) for $49.99
  • A TP-Link Archer C3200 (AC3200) for $119.99

And the deals aren't restricted solely to vendors whose names start with the letter "T" as a quick perusal of DealNews or Techbargains will make clear. What's going on here?

Part of the reason, I suspect (or at least hope), is that consumers are finally wising up to the chronic spec inflation that's long plagued the consumer networking market. Take latest-generation 802.11ac, which regularly receives well-deserved scathing critiques from the folks at SmallNetBuilder. As another SmallNetBuilder tutorial post documents in great detail, all but the lowest-level foundation 802.11ac performance flavors requires one or multiple of the following:

  1. 40 Mhz "multi-channel bonding" in the 2.4 GHz band (which will make you a persona non grata with your neighbors attempting to use that same spectrum swath, due to channel overlap and resultant destructive interference)
  2. Similar (and even wider) "multi-channel bonding" in the 5 GHz band
  3. Simultaneous dual band mode ... even though each LAN client is only able to use one band at a time
  4. Multiple simultaneous 5 GHz radios ... even though each LAN client is only able to tap into one of them
  5. Multiple simultaneous spatial streams, not supported by all LAN clients
  6. Exotic modulation schemes ... again, not supported by all (or even, dare I say, most) LAN clients
  7. Inclusion of the incremental bandwidth potential of the short-range 60 GHz band for emerging 802.11ad-supportive routers ... currently supported by very few (or even, dare I say, any) LAN clients


When you pay big bucks for a router with a big number next to the "AC" designation on the box, only to realize that you're not getting speeds any faster than you were with your old "N" router, might you decide not to repeat your mistake next time? And might you also share your disappointing experiences with family, friends, and co-workers? Just sayin' ...

In (slight) fairness to the router manufacturers, newer system designs (and newer ICs and other components inside those designs) do also tend to deliver a more robust coverage footprint, which can result in higher effective wireless bandwidth at "fringe" regions. But no matter how alien-looking the resultant equipment; see, for example, ASUS' RT-AC5300 below, at some point, a single-router setup won't suffice for full coverage of your premises, whether due to the building's size, the attenuating effects of "chicken wire" wall construction, glass and tile attached to walls, metal appliances in the kitchen, or some combination of these and other factors.


Such is the case at my place. As I've written about before, my house is rectangular in shape and two-level in construction. The router (currently an ASUS RT-N66R; see, even I'm not running 802.11ac yet!) is located downstairs in the furnace room, in the middle of the house, and broadcasts simultaneous 2.4 GHz and 5 GHz beacons along with guest-network equivalents. Upstairs, at either end of the house, are Apple Airport Express units acting as single-band access points and connected to the router over wired Ethernet cable. And also upstairs, in the middle of the house, is a dual-band access point connected to the router over a powerline networking spur to service the back deck.

The combination works passably, but is admittedly something of a pain in the ass. Each wireless broadcast source needs to have a different SSID name, so that I can differentiate them from each other. Wireless-connected clients that roam about the house, such as smartphones, tablets, and laptops, stubbornly continue to cling to a now-weak broadcaster after I move them away from it, versus automatically searching for and reconnecting to a now-stronger alternative. The access points communicate with the router in a "dumb" manner, solely consisting of requesting IP addresses and subsequently shuttling data packets back and forth. And the access points don't communicate with each other, either.

This all changes with mesh networking, a relatively recent development (at least in the consumer networking space), and one that for the first time in a long time promises to bring real value to the upgrade process, thereby resulting in the earlier-exemplified price "crash" for alternative traditional convention networking gear. As yet another SmallNetBuilder article explains, mesh networking is an umbrella term encompassing two different implementations of the concept:

First, let's focus on what's the same about this group of products. All use multiple mini dual-band 802.11ac access points, that can themselves connect via Wi-Fi, to cover more area with higher bandwidth than a conventional single-point router, or at least that's the plan. Three of the five use a mesh architecture, in which APs (aka mesh nodes), can connect to each other to get back to the one node (the "root") that's connected to your modem. The Orbis use a more familar router/extender (aka "satellite") approach, where satellites can connect only to the router.

That particular delineation is pretty much a non-issue in my particular situation, since the router is located in the center of the house, with the access points arrayed around it in a hub-and-spokes arrangement.

Other differentiators between vendors' mesh network alternatives include their degree of support for bonded channels, multiple spatial streams and advanced modulation schemes, all translating into varying peak bandwidth claims. Only a few mesh product options support directly tethered network storage and/or printer resources via USB; similarly, in most cases you'll need to supplement the router or access point with a separate multi-port switch, since the mesh device itself only offers a single Ethernet port intended for LAN clients' use. And in some cases, a distinct 5 GHz "backhaul" radio between access points (and between each access point and the router) is supported, while in other cases, the "backhaul" traffic is merged with normal LAN client traffic (a wired Ethernet option for connecting access points to the router is also offered; that same Ethernet port at the router acts as the WAN connection to the broadband modem).

What makes mesh networking so compelling is the high degree of coordination between the access points and router (as well as each other, at least in some cases, as mentioned earlier). You only need one SSID, for one thing. And various network nodes automatically "hand off" LAN clients to each other as those clients move around the house. Other common consumer-friendly features include automatic channel selection (and regular monitoring/as-necessary reconfiguration) based on overall spectrum availability, and automatic QoS prioritization for specific LAN clients based on sensing their relative bandwidth and response-time needs (video streaming, for example, or online gaming).

Yes, you guessed it ... I took the plunge. I'll be testing two different configurations, actually, both TP-Link-based and Google-branded. I started out by purchasing a (highly rated) Google WiFi three-pack system on sale at Newegg for $244.99. Each mesh node implements AC1200 technology, in part by (intentionally, I assume, for previously-mentioned "friendly neighbor" reasons) only supporting 20 Mhz wide channels in the 2.4 GHz band.


But then I found out that Google had recently added mesh support to its earlier AC1900 OnHub routers, so I bought one of those on sale for $99.99 at the Google Store, supplementing it with two gently used (and less expensive) units acquired on Ebay.


Google doesn't seem very enthused about using OnHubs as mesh access points, for some unknown reason; steps for doing so are documented on their support site, after all, and plenty of users have reported success in doing so. AC1900 OnHubs are also larger than AC1200 Google WiFi units, so for aesthetic reasons I may end up using the latter for the access points, relying on an OnHub as the router. Regardless, stay tuned for my hands-on observations and testing results to come in a future blog post (or, more likely, series).

Read the next part in this series: Mesh networking: setting up Google OnHub-based Wi-Fi


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.


Also see:

Loading comments...

Write a Comment

To comment please Log In

FEATURED RESOURCES