Just how FAST is WiFi 6?

[slo_ncek]

So what router should I order for home use?

[Razor512]

I wish Netgear would lower the price for the Nighthawk AX12 router.

[m2234323]

only  issue issue is those people who woint upgrade  i know some who still on a iphone 4s that only have 4.5  and will be  till it it is dead like old battry or upgrade to a nother cheap 99 on sale phone 

[Razor512]

We will have to wait for benchmarks to see if any enhancements work their way down the line.

Since with most companies, they rarely do performance enhancements to older routers, thus you typically have to wait for newer hardware for someone to take a look at old tech and optimize it. for example. https://www.smallnetbuilder.com/wireless/wireless-features/32512-does-an-ac-router-improve-n-device-performance

 

Another example is in the camera world, Most camera companies never do performance related firmware updates, with the worst offender being Nikon where they only ever do bug fixes ad never performance updates, and Fujifilm being the best, with performance improving significantly over the life of the product.

 

But one example from a company that is scared of firmware updates (Nikon), you will see them use the same AF and metering module 4-5 generations of lower end cameras, but even though thw hardware is the same (completely interchangeable), newer models will have better performance, e.g., an older model may have experience focus shift depending on the color temperature of the lighting, and newer cameras using the same hardware, will have in their firmware some code to compensate for that shift. They could easily port those enhancements over to the older cameras with the same AF and metering module, but they choose not to.

 

The same happens with WiFi routers, and there are chances that the 802.11ax units may improve 802.11ac and probably even 802.11n performance, as all of that code will get another fresh round of testing and optimization.

And as seen if you look at the router charts there, the router and firmware can greatly impact throughput. e.g., compare the Netgear R7800 to the Synology RT2600ac (both use the same hardware).

[Tohnz]

On 5/3/2019 at 9:32 PM, comander said:

I've found that wireless-ac units tend to have a bit better coverage. My guess it's that it comes down to better antenna design and processing (not usually a HUGE difference). With that said more APs is the ideal solution to coverage as you already determined. 

Yep, and that's one of the reasons I did not switch to 802.11ac. As I stated, I don't care to have over 100mbps or low latency in mobile devices ( I just need to put an AP in the one room I use the most to nail it) but I need to have a more uniform coverage when moving around my property. Some walls are even steel reinforced so it's a nightmare for wifi and I know a must live with a multi-AP setup. My fear is that having APs using all the available spectrum and screaming even louder, but still requiring me to use many of them due to thick walls, will just put them working in a more noisy environment and thus experiencing even more dead zones and issues.

On 5/3/2019 at 7:47 PM, Razor512 said:

With the above in mind there are still some benefits to having multiple APs

Nice post Razor, I found it informative. But you missed a couple of things: asymmetric bandwidth and asymmetric power. If one AP is outputting 800mW it will easily be copied by a smartphone with its low gain and low power antenna, but when the smartphone tries to send packets back it will just fail if it's too far (or get near to 90% packet loss at lowest data rate). In that scenario, the device is still connected and the AP seems very strong and can still provide high data-rate, but the device will experience continuous freezing because it cant' send ACKs back or make new requests (the AP can't hear it).

 

More than benefits, it's needed by my very environment have more APs. I must use more than 3 APs, so spanning on channels 1,6,11, overlapping can't be avoided, so I can't ignore the interference they will do each other. So, by contrast, it would be much better putting more weak APs for better coverage than fewer at higher power if clients are low-power rated. (and that's is also the base of 5G coverage)

 

Issues I'm experiencing today are, for example, that my phone stick to the first AP it encounters and if I change the floor or move from a wing to another wing of the house, it doesn't roam to the closest ap. That's is the main reason I want to rig again my network (currently ap's are from different vendors and/or different models, maybe that's the reason roaming sucks, or maybe is happening exactly what described before: the smartphone hears a strong AP, but the AP barely hear the device). Another problem is that when someone in my house uses wifi continuously on a different floor than another, background noise rise up and dead zones enlarge.

 

What I'm looking for is that all those APs talk to each other to negotiate client roaming and avoid cross-talking in areas where they overlap. I don't know if I can achieve it using mesh solutions and I did not yet investigate if wifi6 also has some standard about multi-ap negotiations that can help or not solve these issues.

 

 

[Razor512]

Sadly for sticky clients, Router makers are not really focused on any good solutions, it has been that way for years.

 

For example, we have 802.11 k,v,r to assist in roaming, but in consumer routers,we will rarely see those implemented, thus there is very little on the router side to force clients to a better AP. Roaming at the consumer level is almost entirely directed by the client devices, and most client devices are primarily focused on uninterrupted connectivity, and then will not switch easily.

While it may not be part of a standard, there are ways for them to exercise more control, for example, implement 802.11k,v,r together, and if a client resists a band steer or request to switch, then forcibly disconnect that client and ignore all frames from it, it would be as if the AP was unplugged. (if you have multiple APs in your home, try unplugging one and see how fast a client device will switch).

 

Or imagine if they could develop a way to hand off a session from one AP to another, such that from the perspective of the client device, the AP magically teleported closer, instead of having the client disassociate and then re-authenticate with another AP.

 

 

As for the weak client device, I touched on that topic earlier. Basically most quality APs are able to achieve a massively better receive sensitivity and achieve a lower noise floor internally, as they do not have to compromise in those areas to form factor or power requirement (you are not worried about batteries and screen on time for an access point).

 

Imagine if work was done on the AP side to get them to work together to achieve a result like vmotion on vmware. Though unlike moving a VM, the connection state of a WiFi client in only a few KB of data that is stored in memory. Overall, imagine if client connections could be compartmentalized like that so that client sessions can be moved transparently from the client.

 

[Tohnz]

3 hours ago, comander said:

Some things to note - if you're on 5GHz, wireless-AC will usually be more power efficient on the client side. 

In Italy we have brick walls (my house is very old and I have walls 10 to 50cm thick), 5Ghz won't pass them at all. I've tried and in that spectrum, I'll need one AP in any room . By contrast, I'll benefit more from the 802.11ah at 900Mhz.

 

 

3 hours ago, comander said:

On the 2GHz side, concerns about cochannel interference are often over stated. If you have a line of APs, you can do 1-6-11-1 and be fine - signals drop at a quadratic rate (double the distance and you end up with 1/4th the signal strength). If you have a square, then you can get by by reducing the transmit power on the two APs on the same channel. 

You are right, however, that's is the very tricky part: setting every AP at the right power so overlapping zones don't' get too noisy while not getting dead zones is very hard to achieve, especially if the area is not homogenous as my house is. And consider one thing: in the overlapping zone, a device may receive signal from 2 far AP on the same channel, while those AP don't sense each other: so whenever an AP is granted to transmit to the connected device, there is no guarantee that the other AP starts to transmit to another device at the same time, making the device to temporarily drop all frames; in real life what you experience is that when setting up the network it looks like you cover all the zones, but when people start using it some zones are just covered by noise at a very irregular way, making it difficult to understand if the problem is due to a device sticky connected to the wrong ap, network congestion or whatever.. because just a few seconds later it works just fine... and then drop again suddenly.

 

3 hours ago, comander said:

vkr all help, as does setting a min RSSI but it isn't a magic wand.  

That's is something I will explore better, thank's to remember me about'em. A not long time ago I build a custom firmware (LEDE/LuCI on a TP-Link TL-WA701N/ND v2) for one of the AP to give it a try. I needed to recompile the firmware myself because that AP is cheap with a very low amount of flash. Since I required them to do also EAP (RADIUS) auth I had to tailor what module to include statically and support for v, k and r were left out because there was no space left. As stated, all those AP are planned to be replaced...

 

3 hours ago, comander said:

For what it's worth, all (or at least most) APs being the same brand is generally best practice. Also you'd probably benefit from one additional AP if you have that many dead zones. 

You are right again. My Wifi network has been built over years adding AP where new coverage was required, One of the AP is even just 802.11g  

3 hours ago, comander said:

Also mesh can be hit or miss. Something to keep in mind - it's all on the same channel with most current implementations. 

 

Here's a decent read: https://www.smallnetbuilder.com/basics/wireless-basics/33180-how-to-fix-wi-fi-roaming

Thank's for the link, I'll check it out!!

 

 

[Tohnz]

13 hours ago, Razor512 said:

Sadly for sticky clients, Router makers are not really focused on any good solutions, it has been that way for years.

 

For example, we have 802.11 k,v,r to assist in roaming, but in consumer routers,we will rarely see those implemented,

Exactly, that's the reason I'm using custom firmwares (LuCI and OpenWRT), However, as stated in another post, my old APs are too cheap to allow simultaneously RADIUS auth and 802.11r support (and VLANs, and logging, and bloatware I've wasn't able to strip) and not all the APs I bought can support a custom firmware at all.  I didn't state before, but I'm using it in a mixed way: I use my APs to provide personal access to the internet at home, but I'm also using it for my business SOHO (hence the radius auth and the VLANs).

 

Eventually, I'll change all the APs to new ones with better hardware.. but before I invest so much money and time, I'm must be sure about what I need to achieve my targets. Will be 802.11ac good enough or Wifi6 will be a must, will 802.11r help, are those new mesh solution better, will 802.11ah 800mhz of any help?. there are too many questions I don't have answers yet and I will enjoy any guide that's is not just toward maximum speed but also regards wifi ecosystem in its whole.

[Tohnz]

12 hours ago, gabrielcarvfer said:

That's part of what 5G promised and manufacturers are already working on that.

5G and WIFI are not related to each other.. I'm I wrong? 

[greenhorn]

3 hours ago, Tohnz said:

5G and WIFI are not related to each other.. I'm I wrong? 

Not really, but there are similarities. This is to be expected since some of the challenges they are facing are similar, no need to completely reinvent solutions that work for one of the two if they are applicable to the other one.

Edited May 6, 2019 by greenhorn
typo