Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

navier

macrumors regular
Original poster
Aug 4, 2011
221
230
Germany
The last month I researched for a way to improve the Wi-Fi in my home, as the Apple TimeCapsule won't cut it on the Bedrooms, and found the Wi-Fi Mesh systems.

The idea is good, but the execution not so... (privacy, reliability, etc)

Wouldn't it be great if Apple would release its own Wi-Fi Mesh system?

Now that the HomePod has been released, the team that worked on the Airports that might have been shifted to the HomePod development (my supposition), may go back to make even better Wi-Fi products with modern features like MU-MIMO and auto QoS...

How likely is that we will see such products in the near future?
 
Which Wi-Fi mesh systems did you look at?
The mesh setup I have has been very reliable (fast, no connection issues etc...) and I'm not aware of any privacy issues.
At-least not as far as the way the technology works.
The main issue is usually the loss of bandwidth and added latency with each hop, though if the connections between the mesh access points is fast enough this shouldn't really be noticeable.

Would be interesting to see Apple make one though.


An alternative to mesh might be to use power line adapters. I've had both good and bad (overheating, connection drops, failures) experiences with them, though I've only used Devolo ones.
 
Netgear Orbi: Users report issues with AirPlay not working

Google WiFi: You need a google account, and... it is Skynet... I mean google you know, so privacy it is a issue.

Eero: It needs a permanent connection to Eero central servers. Security issues. Privacy also an issue.

Plumbe: same as Eero.
 
Last edited:
Ah yeah I did consider the Orbi and Google WiFi.

I ended up going with a Ubiquiti UniFi Mesh network. Everything is locally controlled, though there is an optional cloud access if I wanted it. I didn't need to setup any account for it to work, it doesn't require a permanent internet connection and I've had no issues with airplay.

Ubiquiti also have their AmpliFi Mesh network which is aimed more at consumers, though I don't know if that one requires an account or not. From what I've heard it doesn't.
 
The last month I researched for a way to improve the Wi-Fi in my home, as the Apple TimeCapsule won't cut it on the Bedrooms, and found the Wi-Fi Mesh systems.

The idea is good, but the execution not so... (privacy, reliability, etc)

Wouldn't it be great if Apple would release its own Wi-Fi Mesh system?

Now that the HomePod has been released, the team that worked on the Airports that might have been shifted to the HomePod development (my supposition), may go back to make even better Wi-Fi products with modern features like MU-MIMO and auto QoS...

How likely is that we will see such products in the near future?

My bet is on highly likely, HomePod and\or AppleTV as a way to extend the Mesh network from an updated router and/or Home Server I am speculating.

As for your current situation, you can add another Extreme or Express in a remote part of your home to extend the range of the Wi-Fi. But, connect it to the TC via ethernet for best performance. If stringing cable is a challenge, consider MOCA or Powerline adapters to use existing wiring to make the run to the remote location. Newer products in these categories can reach 1Gbps speeds in theory, and have worked really well in my situation.
 
I use 4 Airport Extreme's in my house and move around without drops. Not mesh but roaming and it works great. I have each connected via ethernet to a 5 port switch and each is only used as an access point. My router is standalone. Near the AE I see 400Mbps plus in each direction. The other good thing is the current price of these now, $60 for a super reliable AC router. Like them or not.
 
^The problem will be "legacy" devices like BD-Players, TVs an so on...

Maybe the could use Bluetooth as backhaul "channel" an Wi-Fi for the devices as a bridge solution
 
Maybe the could use Bluetooth as backhaul "channel" an Wi-Fi for the devices as a bridge solution

BT doesn't offer the bandwidth or intelligence needed for the uplinks in Mesh networks.

The solution is to use different 5Ghz channels and antenna\chips. If the radio is connecting clients on channel 36, but uses channel 165 for the uplink, traffic will not collide. Think of this like FM radio, if you are tuned to 107.x, you don't hear signals from 91.x. But, to be truely effective, the uplink should use a different signal path (chip+antenna), so client traffic comes in on one radio, and out on a second. This is effectively how roaming Wi-Fi (Apple's term) works when uplinks are via Ethernet, client Wi-Fi connections are forwarded to the router via Ethernet chip versus retransmitting on the same radio\antenna that received the packet.

The thing that slows Wi-Fi down is collisions. The radios can send or receive a single packet at a time on a given frequency, and there are limits to the effectiveness of probing for a clear channel, so radios simply send and hope for the best, then retransmit until they know the packet was received on the other end. If two packets hit a single radio on the same frequency\channel at the same time, one is dropped and the sender must re-transmit. Separating the signal paths allows each chip to capture or send a packet at the same time and let the device manage routing the packets between the signal paths.

BT and 5Ghz range is about 33 feet give or take depending on obstructions, 5Ghz might be a bit longer range due to stronger radio rx\tx strength but is more susceptible to interference from walls, floors, etc. BT uses 2.4Ghz frequencies, so it suffers the same collision\interference problems that 2.4Ghz Wi-Fi does (the reason 5Ghz was developed), but is a weaker signal with less frequency range (bandwidth) to handle the load. BT bandwidth is something to the tune of 25Mbps versus up to 450Mbps for 2.4Ghz Wi-Fi, and 1Gbps+ for 5Ghz 802.11ac. Another way to think of it is BT is like AM radio, and Wi-Fi like FM. FM offers more bandwidth so separate signals are sent in stereo with meta-data (song name, artist, station name) where AM sends a single mono signal with little data (station name). 802.11ac achieves higher speeds with multiple antenna (MIMO) and beam forming (adapting signal strength based on the distance to the peer). So, 802.11b,g,n (2.4 or 5Ghz) vs 802.11ac (5Ghz) can be thought of a bit like stereo (802.11b,g,n) vs 5.1 (802.11ac).

For Mesh gear to be optimized, it must be able to understand all of the nearby radios, what signal frequencies each is using for both the backhaul and client connections, and be able to calculate the best route at any given time to optimize speed. Consider 4 access points (A - D), with A being the router, and B, C, D being roughly equal distance from the router and each other. If the router has 4 radios (1 for client, 3 for the uplinks to each peer access point), then A-B, A-C, and A-D routes should be equal speeds. But, if D is farther from A, but closer to C and B, the best route to D might be A-B-D or A-C-D. These might also change in short bursts due to traffic, or interference, so being able to adapt to the fastest route will optimize speeds. Again, consider D being farther from A, A-B might have a stream of traffic, so A-C-D might be a faster route, but this could reverse in a second or two when clients connected to C start a stream. Add a 5th or 6th access point and the complexity multiplies.

In pre-mesh access points, they identify a single path to the router, using the same antenna\radio and frequencies as client connections, and alternate\retransmit until they succeed. With 4 Access points and numerous client streams, the traffic starts looking like the Interstate at rush hour with several major accidents. In fact, Apple generally states that Wi-Fi uplinks between a router and access point be limited to one, but more access points can be used effectively if you use Ethernet for the uplink signal paths as Ethernet is better at managing the collisions.

Mesh attempts to adapt to alternate paths based on traffic and speed\distance using multiple radio and frequency paths. Mesh requires bandwidth and intelligence to be able to understand and adapt in order to provide sustained increased service levels. BT is not up to this level of complexity, it is designed for near range communications with little management traffic.
[doublepost=1500478009][/doublepost]One more point, the BT Mesh networks they are referring to in this article are the limited bandwidth BT signals which up to now have been a 1-1 type of communication. BT mesh would allow BT devices to forward signals from peers to extend range, but would not offer the kind of bandwidth that Wi-Fi demands. BT is fine for keyboards, mice, headphones, and even Smart Home device management, but not sufficient for Wi-Fi.
 
^The problem will be "legacy" devices like BD-Players, TVs an so on...

Maybe the could use Bluetooth as backhaul "channel" an Wi-Fi for the devices as a bridge solution

I bow to Techwarrior's superior knowledge. In light of his comments re: bandwidth, contention, etc... One possible way to hybridize BT and Wi-Fi might be to route infrequent, small-scale communications (control of most Smart Home devices, etc.) via BT while routing larger traffic (especially streaming media) via Wi-Fi.

For that matter, with RF reception being spotty by its very nature, a variation on diversity communications makes sense. Mesh offers a form of diversity - choosing the best connection from a variety of devices on the network. However, all nearby devices may suffer from the same poor connection to the router. Giving IoT devices a choice of networks as well as nodes within a network may be advisable - Wi-Fi or BT, whichever is better. Of course, that would make the licensing organizations very happy, and increase the cost of each IoT device, but reliability and ease of use could improve substantially. Give the system the ability to load-balance between networks, and the situation could improve even more.

The BT Mesh announcement describes large-scale BT Mesh networks in commercial buildings, etc. This makes sense. Why tie up large numbers of a Wi-Fi router's IP addresses with devices that communicate relatively infrequently, when that network also has to support an office full of workers? By comparison, any home smaller than Bill Gates' puts a piddling load on a Wi-Fi network.

The more I read about the topic, the more it seems that multiple networks is the necessary solution for now. I understand the appeal of leveraging an existing wifi network for IoT, but considering how marginal many home networks are, dumping far more devices onto that network seems overly ambitious.
 
  • Like
Reactions: techwarrior
I bow to Techwarrior's superior knowledge. In light of his comments re: bandwidth, contention, etc... One possible way to hybridize BT and Wi-Fi might be to route infrequent, small-scale communications (control of most Smart Home devices, etc.) via BT while routing larger traffic (especially streaming media) via Wi-Fi.

For smart home devices (IoT) and other types of device which need mesh, low power but don't need to transfer that much data, currently 802.15.4 is probably a better solution, which runs protocols like Zigbee, DigiMesh (the one I use the most) and Thread. Alternatively there are others like Z-Wave and WeMo. It also has the bonus of being capable of long range (a few miles) depending on the frequency used. An example product would be the Philips Hue range of lightbulbs.
 
  • Like
Reactions: techwarrior
Ubiquiti also have their AmpliFi Mesh network which is aimed more at consumers, though I don't know if that one requires an account or not. From what I've heard it doesn't.

It needs a google or facebook account for remote access.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.