Well.....WiFi 7 is not even ready on OpenWrt now, and Mediatek is announcing the WiFi 8 already.
4 Likes
Wifi 8 isn’t looking very exciting. No actual bandwidth improvements just tweaks to improve latency and reliability, which while welcome aren’t going to be anywhere near as compelling as 6Ghz@320Mhz and MLO has been for Wifi 7. Looks like it might be bringing even less than wifi 6 did, at least there they brought a headline increase in performance at 5Ghz to 1200Mbps.
Depending on how the coordination between access points and beam forming works that could bring about some healthy improvements on mesh and commercial setups. But this area of “seamless roaming as you move through a space” has been chased and “solved” since wifi 6 and they aren’t there yet. Until we get to coordination between all APs in an area including ones we don’t own I suspect we wont see much movement on this problem, other networks will always bring interference.
I think the next big change they should be considering is making use of all the spectrum for all APs. All 60Mhz of 2.4, all 180Mhz of 5Ghz and all 1160Mhz of 6Ghz. Have networks that can utilise the entire lot together but work together with other APs to share spectrum as fairly as possible. We could get peaks in performance approaching 20gbps if we did this but also have it shared with your neighbours in such a way that there is no more setting of channels the APs just detect each other and negotiate a fair share of bandwidth as a requirement of the spec.
One way you handle old setups that have channels is you know they are there so you just remove that spectrum, or contest it and use it at whatever the noise level is at reduced speed but everything else runs without issues. Its the ultimate MLO and getting rid of channels setup that uses all the unlicensed spectrum as fully as possible taking into account geo issues like DFS etc etc. Just always getting you the max you can get in any given second avoiding everything else.
2 Likes
I actually consider the wifi 8 goals of improved latency and reliability more compelling than more throughput from wifi 7. I have no reason for more speed beyond wifi 6, but there are still so many little issues with wifi latency and connectivity, especially taking into account AQL tuning topics. Might as well skip wifi 7 at this point since there is no good OpenWrt device I’m aware of for it yet.
Here is to hoping for a good mt76 wifi 8 device for OpenWrt 30.03 (edit: number symmetry of course there is no branch date 
)
8 Likes
So far that is only a marketing promise, not anything you or me (or larger numbers of real-world users) have independently confirmed, just wait and see what really gets delivered (and consider the huge delay between commercial availability and OpenWrt support).
3 Likes
I did jokingly say 30.03 which would be 4.5 years away after the RC process. Yea wouldn’t buy anything until fully OpenWrt supported anyway. Time will tell.
You aren't far off. wifi8 is only being announced now - it'll take another ~year until you can actually buy it on the shelves, for top dollar - and another 2+ years until it will tickle down into cheap devices, expect another 2-3 years of development after devices become available for-cheap to mere mortals (which includes the OpenWrt developers, they generally don't get early hardware access, nor could afford the first prosumer devices entering the market).
2 Likes
Yes I agree with this, but for those who haven't get WiFi 7 yet (like me, I don't even have WiFi 6E) it makes sense to go straight to WiFi 8.
But latency/reliability improvements are essential for gamers, I know hard wire is always the best, but there are many situations that people have to use WiFi and this will greatly help.
I hope there would be some early development boards like BPI-R4 so that the devs can start looking at it soon.
1 Like
Not really, we already have basic wifi7 support right now (filogic, ipq957x isn't impossible either), it should stabilize within this year - and devices hopefully become affordable.
wifi8 devices will take another year to hit the shelves, initially in the upper mid-3-figure price range - and OpenWrt support is still years away (2030 is NOT unrealistic).
wifi6 devices are cheap(ish) and plenty (including supported ones) right now, you can buy them today without the fear of missing out. wifi 6e is 'only' a question of pricing.
Skipping wifi6/7 would only make sense, if you could buy OpenWrt supported devices for reasonable prices (150-200 EUR/ USD) anytime soon, but that's still years+ out. And by the times that happens, wifi9 will be all the craze in the tech news.
I think you haven't seen this thread then
60W would make this tech DOA. But I don’t really get how power would increase that much. The power per radio has remained steady and it’s not like FCC will allow higher power outputs. For example my MT6000 (2 years old) draws very similar power to my WRT32X (10 years old) also thanks to new lithography. 2.5Gbit draws a bit more but even those PHY power have come down. Also an Intel N150 box can idle at 12W (20W loaded) now thanks to the tickless kernel. My point is it’s probably not 60W that’s just the power supply. Provided it uses better lithography and has an efficient design even 20W would be absurd.
3 Likes
- "will cater to premium and flagship devices"
- All AP must be upgraded to wifi8 + have wifi8 devices
- broadcom and qualcomm will continue to stomp out filogic models
- zero, easy to understand and easily quantifiable marketing checkboxes
For consumers, it will have worse adoption than wifi7.
For openwrt users, the vast majority are at wifi6. 6e is skippable. So you either slowly adopt wifi7 or wait 5-7 years until 8 arrives with stable support... assuming filogic models exist. The only wifi7 model out with okay support is the BT8 and it appears very few have bought it. Purchases = votes to support filogic models.
Its still a $300 ish device which puts it well out of range of what most people expect for a router. Its also pretty compromised for the high end being limited to 2.5gbps ethernet. The fact that now the high end devices in wifi 7 are $600+ is really bad and its got to be limiting the volumes substantially.
I feel like generally network kit is just getting expensive, they have deemed anything more than 1gbps as premium now and charging appropriately.
1 Like
I'm seeing $150ish in EU and USA.
Asus did confirm a hopefully cheaper filogic wifi7 model at CES for Q1 launch:
RT-AX59U --> PRT-BE59 PRT-BE5000
- BT8/BT6/BE14000 - $150-$250
- Banana Pi BPI-R4 - $150
- should be filogic RT-BE5000 - $150-$200
- mythical Openwrt Two - $250 target price--if it ever exists.
- mythical Flint2 replacement ("Flint 4" looks $$$$)
The future for wifi8 filogic looks grim if Asus dumps them based on terrible sales.
Asus ZenWiFi BT7 is coming, but it's using IPQ5322.
RT-BE90U / RT-BE9400 will soon be released, but it is basically a rebrand of TUF-BE9400 (IPQ5322) based on FCC MSQ-RTBE8T00 C2PC documents.
GS-BE7200X (MT7988D) FCC MSQ-RTBE9B00 documents were released couple days ago so it should be coming into the US / Canada market at the very least, but it’s a dual-band (2.4/5GHz) and I think it’s going to be pricey considering the “ROG” tax.
TP-Link Archer BE260 or BE5000 is using MT7987 based on FCC 2BH7FBE5000 internal photos. It’s currently available at Amazon for ~$100.
Archer BE450 (MT7988D) seems to be in the Asia market only.
ipq957x has at least some kind of basic development (source-only), ipq53xx has none so far (doesn't mean that it will never happen, just that it will take considerable time, well in excess of a year).
FLINT 4.
1 Like
Should make it a dual 10g + 4x 2.5G device like BPi R4 Pro. And keep it 4x4 on all bands. And… get rid of the touch screen. Waste of $$$ for something that usually sits in a closet or under a desk.
1 Like
Are there any consumer-oriented routers that actually support MLO out of the box? That is proper multi-radio and not just marketing claims. Last I checked that wasn't the case, which makes it seem like even Wi-Fi 7's biggest promise isn't there yet:
Otherwise these seem like more routers that are no-better than AX. Mostly throwing shade at Gl.Inet to be honest 
1 Like
I told them to make filogic + 1-2 10G + 8 2.5G port without wireless. Goes against the "w" in openwrt, but I wish wired gateways became a thing in the consumer market.
1 Like
Ah, MLO why not have a single station hog airtime in all wifi bands simultaneously, primarily in the quest to deliver higher best case throughputs in artificial capacity tests.... at a stage where most interactive applications fail to even saturate 100 Mbps...
Sure the smoother handovers between bands are nifty, but honestly for me a luxury problem...
Personally I think OFDM in wifi6 was a bigger improvement, albeit one that will payoff most in the future when more devices are actually ofdm capable. But then these assessments are clearly subjective...
P.S.: I have been told that OFDM is optional in WiFi6 (and hence rare) but became mandatory in WiFi7, I have also been told that OFDM really noticeably helps with high station density, so there might be a good reason for WiFi7 after all...
2 Likes