Ruckus Wireless as an Unifi alternative?

[itronin]

Is it really new?
I think 20 bucks to get an "enterprise" grade ap and access to a device running unleashed (it's awesome compared to consumer aps & their firmware/os) is okay.

if this is the "woot deal" that was posted then they are indeed new. You can read about the deal here.

Its tiny for an enterprise grade AP.

it will take poe *or* an AC adapter and those can be had cheap like $1USD (the bay) to < $10 (AMZN).

It can run Ruckus Unleashed. Has great antenna tech. did I mention its tiny?

the only negative from my perspective is that it can't do wireless mesh.
If you want that and a pure AP mesh solution then my advice is to get some used R710's. The ruckus mesh using R710's works better than any of the consumer/prosumer stuff I've tried in challenging locations (Asus, Orbi, UBQ etc.)

NB: That didn't stop my from buying a couple just to have around and plug in as necessary (back porch, garage etc).

Overall rating: "Tastes Great - Less Filling".

 

[gigi-biji]

if this is the "woot deal" that was posted then they are indeed new. You can read about the deal here.

Its tiny for an enterprise grade AP.

it will take poe *or* an AC adapter and those can be had cheap like $1USD (the bay) to < $10 (AMZN).

It can run Ruckus Unleashed. Has great antenna tech. did I mention its tiny?

the only negative from my perspective is that it can't do wireless mesh.
If you want that and a pure AP mesh solution then my advice is to get some used R710's. The ruckus mesh using R710's works better than any of the consumer/prosumer stuff I've tried in challenging locations (Asus, Orbi, UBQ etc.)

NB: That didn't stop my from buying a couple just to have around and plug in as necessary (back porch, garage etc).

Overall rating: "Tastes Great - Less Filling".

Thanks, yes, from Woot and I just pulled the trigger. I got POE+ Ethernet everywhere, wouldn't even cross my mind to do wireless mesh even if it was supported.

 

[blunden]

Don't think about MIMO 2x2, nothing really can use 4x4 or something like it.

That's not really how that works. Yes, most clients are only 2x2, but a 4x4 AP can then talk to multiple 2x2 clients at the same time so it still matters.

For home use and slower clients you probably won't notice the difference unless you have a large number of connected clients, but 4x4 is not useless just because you don't have 4x4 clients.

 

[namezys]

That's not really how that works. Yes, most clients are only 2x2, but a 4x4 AP can then talk to multiple 2x2 clients at the same time so it still matters.

Sorry, it doesn't work in this way. I was wondering how it can be real because my scientific knowledge said different things and got proof.

MIMO 4x (send 4 radio stream) can not interact with two MIMO 2x (receive 2 radio stream) clients because client should be able to rebuild 4 signals but can only 2.

However, MIMO 4x can work if AP can form beam in such way that they will not be a interference in client point, usually it work on long way paths.
Also it gives ability to build beams more precise.

It is not about h320, it is pretty simple in terms of beam patterns but it is enough for home usage.

 

[blunden]

Sorry, it doesn't work in this way. I was wondering how it can be real because my scientific knowledge said different things and got proof.

MIMO 4x (send 4 radio stream) can not interact with two MIMO 2x (receive 2 radio stream) clients because client should be able to rebuild 4 signals but can only 2.

However, MIMO 4x can work if AP can form beam in such way that they will not be a interference in client point, usually it work on long way paths.
Also it gives ability to build beams more precise.

It is not about h320, it is pretty simple in terms of beam patterns but it is enough for home usage.

I'm not at all convinced by that, sorry.

Not sure what you're even trying to say with "
because client should be able to rebuild 4 signals but can only 2."? Why should a single client have to rebuild 4 signals?

 

[namezys]

I'm not at all convinced by that, sorry.

Not sure what you're even trying to say with "
because client should be able to rebuild 4 signals but can only 2."? Why should a single client have to rebuild 4 signals?

I can try to explain even I'm sure that I will use correct English words.

Assumption:
* AP (access point) has MO x4 (multi output)
* ST (station) has MI x2(multi input)
* Data is going from AP to ST at the same frequency channel.

In this case:
* AP has 4 antennas that are located with some gap (let's call them A B C D)
* ST has 2 antennas that does not located in same point (lets call them X Y)

In this case path from any AP antenna to any ST antenna is different.
In other words, path A-X is different from path A-Y.

Each AP antenna send independent signals: Sa, Sb, Sc, Sd
Each ST antenna receive sum of signals from all antennas of AP: Sx, Sy

a1 * Sa + b1 * Sb + c1 * Sc + d1 * Sd = S1

a2 * Sa + b2 * Sb + c2 * Sc + d2 * Sd = S2

The goal is to reconstruct Sa, Sb, Sc, Sd (even we don't need all of them) using known Sx, Sy.
During learning ST is calculated path lengths. It will static coefficients in linear equations.

Finally, we will get 2 linear equations with 4 unknown signals. It does not have unique solution there for we can not reconstruct signals Sa, Sb, Sc, Sd. Even ST needs only 2 signals, there is no unique solution for it (in general).

In case of beam forming we can assume that coefficients for some signals is zero (in other words, these signals are very weak for ST). Therefore we can solve it and reconstruct source signal. However, it's very complicated to make such beam forming.

 

[gigi-biji]

The Wootman (Amazon delivery really) dropped them of today. I configured one and doing a basic test, placed next to my Asus RT-AC68 in the lower level of the house. The Ruckus is placed with labels down on a table (so it is beaming up) and I am on the second floor (semi open floor plan). Did dozens of iperf3 local speed tests and the ASUS is destroying the single r320 on 2,4ghz. ASUS is consistenly coming up 2-3 times faster at same location of the house.


That just early tests, I have barely touched the configurations, but if anyone has any tips (Tx power settings?) - please share. It's been a long day, so will continue messing with it tomorrow.

 

[fohdeesha]

The Wootman (Amazon delivery really) dropped them of today. I configured one and doing a basic test, placed next to my Asus RT-AC68 in the lower level of the house. The Ruckus is placed with labels down on a table (so it is beaming up) and I am on the second floor (semi open floor plan). Did dozens of iperf3 local speed tests and the ASUS is destroying the single r320 on 2,4ghz. ASUS is consistenly coming up 2-3 times faster at same location of the house.


That just early tests, I have barely touched the configurations, but if anyone has any tips (Tx power settings?) - please share. It's been a long day, so will continue messing with it tomorrow.

are you running them at the same time? turn the other one off when testing one of them, or theyre going to stomp all over eachother's spectrum, and asus being consumer stuff probably has TX cranked to a static max instead of auto like the ruckus. SOunds like it's also using 40mhz or even 80mhz channel width instead of 20 like the ruckus, which will also stomp all over its spectrum and make it slow as hell

 

[gigi-biji]

are you running them at the same time? turn the other one off when testing one of them, or theyre going to stomp all over eachother's spectrum, and asus being consumer stuff probably has TX cranked to a static max instead of auto like the ruckus. SOunds like it's also using 40mhz or even 80mhz channel width instead of 20 like the ruckus, which will also stomp all over its spectrum and make it slow as hell

Both are on, but I am testing one at a time and there is no heavy traffic on existing ASUS network. Asus Ch6, 20mhz, Ruckus ch1, 20 mhz. I will be doing more tomorrow, but so far the difference is dramatic.

 

[gigi-biji]

are you running them at the same time? turn the other one off when testing one of them, or theyre going to stomp all over eachother's spectrum, and asus being consumer stuff probably has TX cranked to a static max instead of auto like the ruckus. SOunds like it's also using 40mhz or even 80mhz channel width instead of 20 like the ruckus, which will also stomp all over its spectrum and make it slow as hell

I don't even see settings for Tx on the Ruckus. In fact, I don't see any tweaking for the radio at all. Hopefully I am just tired.

 

[fohdeesha]

I don't even see settings for Tx on the Ruckus. In fact, I don't see any tweaking for the radio at all. Hopefully I am just tired.

you're tired

 

[gigi-biji]

you're tired

View attachment 43102

Thanks, found it. So there are radio options at 3 different groups of menus (WiFi, AP, and Administration). I set 2.4ghz on max power. Unfortunately it still sucks, massively. And the r320 is on very clean channel 1, ASUS on more congested channel 6.

Some observations (2.4ghz only, single AP):
Max link speed of ASUS 5ft away: 144Mbps ; Ruckus - 130Mbps
But the more I move away from them, the bigger the delta.dBm also drops, Ruckus is trailing 3-9 dBm lower.A hallway and a room away (open door), and Asus link speed is still 100Mbps, while Ruckus drops to 52, 26 and even 19Mbps.

I do remember reading awhile back that RT-AC68 was one of the last routers/wap with the strong radios. Maybe there was a power limit change 10 years ago, I don't know ...

I am not able to upload screenshots yet, so here is a co-paste of some iperf3 results (Ruckus is test# 80 on top):

Server listening on 5201 (test #80)
-----------------------------------------------------------
Accepted connection from 192.168.9.42, port 61781
[ 5] local 192.168.9.99 port 5201 connected to 192.168.9.42 port 61782
[ ID] Interval Transfer Bitrate
[ 5] 0.00-1.01 sec 256 KBytes 2.08 Mbits/sec
[ 5] 1.01-2.01 sec 512 KBytes 4.20 Mbits/sec
[ 5] 2.01-3.02 sec 512 KBytes 4.17 Mbits/sec
[ 5] 3.02-4.01 sec 512 KBytes 4.23 Mbits/sec
[ 5] 4.01-5.01 sec 384 KBytes 3.15 Mbits/sec
[ 5] 5.01-6.01 sec 512 KBytes 4.16 Mbits/sec
[ 5] 6.01-7.01 sec 384 KBytes 3.16 Mbits/sec
[ 5] 7.01-8.00 sec 128 KBytes 1.06 Mbits/sec
[ 5] 8.00-9.02 sec 128 KBytes 1.04 Mbits/sec
[ 5] 9.02-10.01 sec 0.00 Bytes 0.00 bits/sec
[ 5] 10.01-10.95 sec 0.00 Bytes 0.00 bits/sec
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval Transfer Bitrate
[ 5] 0.00-10.95 sec 3.25 MBytes 2.49 Mbits/sec receiver
-----------------------------------------------------------
Server listening on 5201 (test #81)
-----------------------------------------------------------
Accepted connection from 192.168.9.42, port 61916
[ 5] local 192.168.9.99 port 5201 connected to 192.168.9.42 port 61917
[ ID] Interval Transfer Bitrate
[ 5] 0.00-1.01 sec 2.00 MBytes 16.7 Mbits/sec
[ 5] 1.01-2.00 sec 2.00 MBytes 16.8 Mbits/sec
[ 5] 2.00-3.00 sec 3.25 MBytes 27.3 Mbits/sec
[ 5] 3.00-4.01 sec 1.38 MBytes 11.4 Mbits/sec
[ 5] 4.01-5.01 sec 1.50 MBytes 12.6 Mbits/sec
[ 5] 5.01-6.00 sec 2.25 MBytes 19.0 Mbits/sec
[ 5] 6.00-7.00 sec 2.50 MBytes 21.0 Mbits/sec
[ 5] 7.00-8.01 sec 2.00 MBytes 16.5 Mbits/sec
[ 5] 8.01-9.01 sec 1.88 MBytes 15.8 Mbits/sec
[ 5] 9.01-10.01 sec 1.62 MBytes 13.7 Mbits/sec
[ 5] 10.01-10.06 sec 256 KBytes 35.2 Mbits/sec
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval Transfer Bitrate
[ 5] 0.00-10.06 sec 20.6 MBytes 17.2 Mbits/sec receiver
-----------------------------------------------------------

 

[delamonpansie]

I want to share a word of warning about new firmware: around 200.15 something was "improved" and I see speed capped at about 300-500Mbit/sec.
What's worse, traffic on 80 port is penalized even more.

From a quick debug I can tell that something is not ok with a driver/kernel. It shows abnormal usage at soft irq during a test. Test is a single client fetching large file http on port 800. The system is R650 running latest 200.17, but I observer the same problem with any firmware starting from 200.15.

Mem: 526208K used, 264440K free, 0K shrd, 515040K buff, 515091K cached
CPU0:  0.0% usr  0.0% sys  0.0% nic 78.3% idle  0.0% io  0.0% irq 21.6% sirq
CPU1:  0.0% usr  0.9% sys  0.0% nic 99.0% idle  0.0% io  0.0% irq  0.0% sirq
CPU2:  0.0% usr  0.0% sys  0.0% nic  0.0% idle  0.0% io  0.0% irq  100% sirq
CPU3:  0.0% usr  3.8% sys  0.0% nic 96.1% idle  0.0% io  0.0% irq  0.0% sirq
Load average: 1.21 0.64 0.25 4/183 5053
  PID  PPID USER     STAT   VSZ %MEM CPU %CPU COMMAND
   18     2 root     RW       0  0.0   2 20.9 [ksoftirqd/2]
31836     2 root     RW       0  0.0   3  0.7 [kworker/u8:5]
 4943  4942 root     R     1224  0.1   3  0.2 top
  676     2 root     SW<      0  0.0   3  0.2 [scheduler_threa]
    3     2 root     SW       0  0.0   0  0.2 [ksoftirqd/0]
 1675     1 root     S    88032 11.0   0  0.0 /usr/sbin/statd -D
 1619     1 root     S    87568 11.0   0  0.0 /usr/sbin/avpd -D
 1685     1 root     S    81232 10.2   1  0.0 /usr/sbin/qm_dpi -D -n 1

As you can see, one core is fully swamped with softirq load, leaving no breathing room.

compare to 200.14

Mem: 319380K used, 475032K free, 0K shrd, 0K buff, 3825K cached
CPU0:   0% usr   0% sys   0% nic  88% idle   0% io   0% irq  10% sirq
CPU1:   0% usr   0% sys   0% nic  99% idle   0% io   0% irq   0% sirq
CPU2:   0% usr   0% sys   0% nic  99% idle   0% io   0% irq   0% sirq
CPU3:   0% usr   0% sys   0% nic  98% idle   0% io   0% irq   0% sirq
Load average: 0.00 0.00 0.00 3/183 4980
  PID  PPID USER     STAT   VSZ %MEM CPU %CPU COMMAND
  441     2 root     SW<      0   0%   0   0% [scheduler_threa]
  315   160 root     S <   5944   1%   0   0% /usr/sbin/timer start
  769     1 root     S     5924   1%   0   0% /usr/sbin/channelfly -i wifi0 -q
  620     1 root     S     2652   0%   0   0% /usr/bin/mosquitto -c /tmp/mosquitto.conf
 3856  3855 root     R     1196   0%   0   0% top
    7     2 root     SW       0   0%   0   0% [rcu_sched]
    3     2 root     SW       0   0%   0   0% [ksoftirqd/0]
  929     1 root     S    34684   4%   0   0% /usr/sbin/qm_dpi -D -n 1
  898     1 root     S    19444   2%   0   0% /usr/sbin/uf_agent
  720     1 root     S    18288   2%   0   0% /usr/sbin/apmgr -r start
 1428     1 root     S    14872   2%   0   0% /usr/bin/stainfod

The system is barely loaded at all.

This is an unleashed network with two APs, test was done on non-master. Both version were configured exactly the same.

 

[fohdeesha]

Thanks, found it. So there are radio options at 3 different groups of menus (WiFi, AP, and Administration). I set 2.4ghz on max power. Unfortunately it still sucks, massively. And the r320 is on very clean channel 1, ASUS on more congested channel 6.

Some observations (2.4ghz only, single AP):
Max link speed of ASUS 5ft away: 144Mbps ; Ruckus - 130Mbps
But the more I move away from them, the bigger the delta.dBm also drops, Ruckus is trailing 3-9 dBm lower.A hallway and a room away (open door), and Asus link speed is still 100Mbps, while Ruckus drops to 52, 26 and even 19Mbps.

I do remember reading awhile back that RT-AC68 was one of the last routers/wap with the strong radios. Maybe there was a power limit change 10 years ago, I don't know ...

I am not able to upload screenshots yet, so here is a co-paste of some iperf3 results (Ruckus is test# 80 on top):

Server listening on 5201 (test #80)
-----------------------------------------------------------
Accepted connection from 192.168.9.42, port 61781
[ 5] local 192.168.9.99 port 5201 connected to 192.168.9.42 port 61782
[ ID] Interval Transfer Bitrate
[ 5] 0.00-1.01 sec 256 KBytes 2.08 Mbits/sec
[ 5] 1.01-2.01 sec 512 KBytes 4.20 Mbits/sec
[ 5] 2.01-3.02 sec 512 KBytes 4.17 Mbits/sec
[ 5] 3.02-4.01 sec 512 KBytes 4.23 Mbits/sec
[ 5] 4.01-5.01 sec 384 KBytes 3.15 Mbits/sec
[ 5] 5.01-6.01 sec 512 KBytes 4.16 Mbits/sec
[ 5] 6.01-7.01 sec 384 KBytes 3.16 Mbits/sec
[ 5] 7.01-8.00 sec 128 KBytes 1.06 Mbits/sec
[ 5] 8.00-9.02 sec 128 KBytes 1.04 Mbits/sec
[ 5] 9.02-10.01 sec 0.00 Bytes 0.00 bits/sec
[ 5] 10.01-10.95 sec 0.00 Bytes 0.00 bits/sec
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval Transfer Bitrate
[ 5] 0.00-10.95 sec 3.25 MBytes 2.49 Mbits/sec receiver
-----------------------------------------------------------
Server listening on 5201 (test #81)
-----------------------------------------------------------
Accepted connection from 192.168.9.42, port 61916
[ 5] local 192.168.9.99 port 5201 connected to 192.168.9.42 port 61917
[ ID] Interval Transfer Bitrate
[ 5] 0.00-1.01 sec 2.00 MBytes 16.7 Mbits/sec
[ 5] 1.01-2.00 sec 2.00 MBytes 16.8 Mbits/sec
[ 5] 2.00-3.00 sec 3.25 MBytes 27.3 Mbits/sec
[ 5] 3.00-4.01 sec 1.38 MBytes 11.4 Mbits/sec
[ 5] 4.01-5.01 sec 1.50 MBytes 12.6 Mbits/sec
[ 5] 5.01-6.00 sec 2.25 MBytes 19.0 Mbits/sec
[ 5] 6.00-7.00 sec 2.50 MBytes 21.0 Mbits/sec
[ 5] 7.00-8.01 sec 2.00 MBytes 16.5 Mbits/sec
[ 5] 8.01-9.01 sec 1.88 MBytes 15.8 Mbits/sec
[ 5] 9.01-10.01 sec 1.62 MBytes 13.7 Mbits/sec
[ 5] 10.01-10.06 sec 256 KBytes 35.2 Mbits/sec
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval Transfer Bitrate
[ 5] 0.00-10.06 sec 20.6 MBytes 17.2 Mbits/sec receiver
-----------------------------------------------------------

Do you have both just sitting on a table? The Asus antenna pattern was designed to do so, the ruckus is designed for the complete opposite, it's meant to be upside down on a ceiling with purposefully null zones behind it to not interfere with APs on floors above

 

[gigi-biji]

Do you have both just sitting on a table? The Asus antenna pattern was designed to do so, the ruckus is designed for the complete opposite, it's meant to be upside down on a ceiling with purposefully null zones behind it to not interfere with APs on floors above

I already said that the Ruckus was sitting on the table with labels facing down (signal beaming up) on a semi-open floor below the testing area.

This morning I moved it up to second floor and mounted on ceiling. One thing is 100% clear at this point:
The RT-AC68U is destroying a single r320 with range and speed on the 2.4ghz channels. The range difference on 5G is not as big, but ASUS still has the edge. However, the R320 is faster on 5G but only in close proximity (same room or so). To cover the existing ASUS range, I will need 3x R320s and that is ok, because it will come with the benefit of much faster(5G) overall speeds thru out the home (Only 1400 sq ft per level, 2 levels, and 4000 sq ft back yard that I am trying to cover with a wap beaming at it).

Now, the big question is, has roaming technology evolved since last time I tried multi WAPs (10 years ago). I hocked second R320, the first one picked it up and auto provisioned it. NICE!

But my phone (Android) has been sitting 3 ft away from the second WAP for 30 minutes, and yet it is still associated with the first WAP which only has -70dBm RSSI. I am guessing out of box roaming settings are not right, any recommendations at what should I be looking at?

 

[tgl]

But my phone (Android) has been sitting 3 ft away from the second WAP for 30 minutes, and yet it is still associated with the first WAP which only has -70dBm SNR. I am guessing out of box roaming settings are not right, any recommendations at what should I be looking at?

Wifi roaming decisions are always driven by the clients, not the APs. I'd venture that your phone thinks -70dBm is Good Enough(TM) and so it's not looking to roam. (Apple documents their roaming rules here. I don't know if there's an equivalent document for Android, or whether it's the same across all their manufacturers.) The thing you usually do when setting up a multi-AP system is to turn the APs' transmit power down enough that the clients will roam at the points where you want them to. This is totally unlike the situation with a single-AP system, where you might as well let it blast away at max Tx power, because either the clients can connect to it or not. In a multi-AP system it's very rare that max power is the right setting.

Also, make sure the APs have 802.11k/v signaling enabled. To the extent that a phone is willing to do roaming smarter than "look for another AP once the signal gets bad enough", it's 802.11k/v information about neighboring APs that it will want.

 

[gigi-biji]

Wifi roaming decisions are always driven by the clients, not the APs. I'd venture that your phone thinks -70dBm is Good Enough(TM) and so it's not looking to roam. (Apple documents their roaming rules here. I don't know if there's an equivalent document for Android, or whether it's the same across all their manufacturers.) The thing you usually do when setting up a multi-AP system is to turn the APs' transmit power down enough that the clients will roam at the points where you want them to. This is totally unlike the situation with a single-AP system, where you might as well let it blast away at max Tx power, because either the clients can connect to it or not. In a multi-AP system it's very rare that max power is the right setting.

Also, make sure the APs have 802.11k/v signaling enabled. To the extent that a phone is willing to do roaming smarter than "look for another AP once the signal gets bad enough", it's 802.11k/v information about neighboring APs that it will want.

Kind of scared me for a moment that it is 2025 and we are still relaying on clients [only] to roam. Looks like 802.11k/v/r are now more broadly adopted, so should be what I am looking for. Thanks for your help!

I am going to keep my RT-AC68U around, in case the 3x r320 endeavor doesn't work. I always suspected it was special, now I appreciate how good it's radio actually is. If I have to turn down the r320 Tx power, I will probably need 5x R320s, which is not something I am going to be OK with.

 

[LodeRunner]

Kind of scared me for a moment that it is 2025 and we are still relaying on clients [only] to roam. Looks like 802.11k/v/r are now more broadly adopted, so should be what I am looking for. Thanks for your help!

I am going to keep my RT-AC68U around, in case the 3x r320 endeavor doesn't work. I always suspected it was special, now I appreciate how good it's radio actually is. If I have to turn down the r320 Tx power, I will probably need 5x R320s, which is not something I am going to be OK with.

The AP side can force a roam (by kicking the client) if you have min RSSI settings, but you’d probably want to do a radio survey before using that.

 

[tgl]

Min RSSI is almost never a good thing to use. The client has no idea why it was kicked and is perfectly likely to try to reconnect to the same AP. (It might get the point after being booted several times, or it might not.) Even in the best case, it'll be far from a seamless roaming experience because the client will have to do a new scan for candidate APs after it's already lost service.

802.11k/v offers a much more plausible answer, but you do need equipment that supports it --- clients and APs both --- and the APs have to be configured to believe that they are part of the same network. IME it's not enough that they are broadcasting the same SSID name; they have to be run from a common configuration/control source or they won't do it. I don't have any personal experience with Ruckus so I'm not sure how that statement translates into their configuration setup.

 

[kpfleming]

IME it's not enough that they are broadcasting the same SSID name; they have to be run from a common configuration/control source or they won't do it. I don't have any personal experience with Ruckus so I'm not sure how that statement translates into their configuration setup.

Yes, this is quite important, and is a big reason why 'WiFi extenders' often fail to provide any value (I just helped a neighbor deal with this). The 'extender' just rebroadcasts the same SSID, but is otherwise not integrated with the existing APs, so the APs and extended cannot assist the WiFi clients in making decisions about where to associate.

The same will happen with independent APs broadcasting the same SSID; they aren't aware of each other.

When a configuration with multiple integrated APs is used (for Ruckus that's Unleashed, SmartZone, etc.) then the 'controller' will help the APs configure themselves to minimize interference, maximize coverage, and assist clients as they roam around the coverage area.