Community discussions

 
n21roadie
Forum Guru
Forum Guru
Posts: 1886
Joined: Fri Aug 07, 2009 10:36 pm
Location: Limerick,Ireland

Re: ARM devices and NV2 protocol

Thu Jan 17, 2019 11:48 am

..........................................
2)Pure 802.11N +rts/cts @20Mhz 90-20 mbit/s sometimes UPLOAD rates are falling to 6Mbps latency spikes. Some clients lose 90% packet loss (sector 20 clients)

NV2 works great for 20Mhz channel it does not scale correctly for 40/60/80Mhz. For good TCP/IP one session speed need rly good SNR from AP side (collocation problem).

More about RTS/CTS are here https://www.sciencedirect.com/science/a ... 0715000129
Very high packet loss means that the antenna has a big issue with interference, NV2 is much better at dealing with interference,
All of which would indicate that the performance/design of a antenna is more important that the wireless protocol used, especially when throughput is increased as any antenna
imperfections are magnified?
 
ste
Forum Guru
Forum Guru
Posts: 1786
Joined: Sun Feb 13, 2005 11:21 pm

Re: ARM devices and NV2 protocol

Thu Jan 17, 2019 11:54 am

I think these could easily been avoided all by indicated in the spec sheet:

Model:
xxxxxxx
Transfer speed via wireless link ( ideal condition )
802 - 400mbps both directions TCP
nv2 -100mbps both directions TCP
nsextreme - 120mbps both directions TCP

something like that. Then the customers would have an informed decision basing on the information given and there will be no threads like this as it would simply be dismissed as " Slower speeds are due to the protocol overhead or it was already indicated in the spec sheets "
Doing this would be a vendor promise which could not be guarateed.
 
2jarek
Member Candidate
Member Candidate
Posts: 108
Joined: Thu May 17, 2007 3:28 pm
Location: Poland

Re: ARM devices and NV2 protocol

Thu Jan 17, 2019 12:21 pm

@n21roadie no packet los because too much collisions from hide nodes UPLOAD rate fall at 6mbps for worse signal station. Clean spectrum RTS/CTS no help very much. Just try BT 1 tcp sesion from 20 nodes simultaneously for 802.11 +RTS/CTS @ change for NV2.
+big ACK problem for pure 802.11N/AC under heavy load & saturation, NV2 works much much better @20Mhz
 
WirelessRudy
Forum Guru
Forum Guru
Posts: 3074
Joined: Tue Aug 08, 2006 5:54 pm
Location: Spain

Re: ARM devices and NV2 protocol

Thu Jan 17, 2019 3:04 pm

So here we go again.. (good! The more discussion the better! :D )

Package loss might be less in NV2, but imho when you have more then double the speed available that is not so much of an issue. The package simply gets resend (tcp) and that's it.

If I have a sector with 10, 20 or even 30+ clients associated, I still see relative little usage in general. No matter when I look, any AP that needs to deliver more then 30Mbps to its clients for a prolonged time is a rare event. Sometimes when a couple of clients open something at the same time I see a spike of 30-40Mbps, sometimes into 50...
So yes, running BT from 6 or more clients then NV2 will probably give each a better share with less package losses

But for the day to day practice it is good enough when a clients request gets handled by the highest possible speed to free the spectrum. When running csma with properly configure rts/cts I see for traffic to only 2 or 3 clients hardly any package losses. The new csma+rts/cts protocol in 'ac' works pretty fine.
I can run from two clients a BT test and both have 90-100Mbps to a server in the LAN that halves if both are tested at the same time. And 1/3 roughly when 3 tests.. But is still much more then the same tests in NV2.

To stay in pace with the competition offering 30 or 50Mbps packages we need to go with it and offer the same. (It's always "up to"). With NV2 I cannot offer that. Not even with an AP that has only 6 clients.... 30 is max. run a second and it collapse...
With csma I can run at least two at almost 50 at the same time. Meaning that if we install, or a single client runs a speedtest, we get what we sell... 30 or 50Mbps....
Off course not when 2 or 3 or more clients at the same time hit the same www.speedtest.net button! But when happens this?

We have a normal client base with facebook, skype, Whatsapp use and video cameras and IPTV/Netflix etc. and occasional downloads. In NV2 we get complaints from users they are not getting their speeds. They usually start to test when something already has gone wrong, like freezing picture of broken IPTV streams...
In csma we get no complaints (so far)
Even several IPTV streams over csma work fine. You'll see the characteristic traffic (high peaks of download followed by pauzes) from serveral clients at the same time and all have smooth HD TV over the csma AP. When the same happens with NV2 the network gets saturated much faster and the complaints follow...

(At 100Mbps air capacity in the csma P2MP network an IPTV stream can use short bursts of high speed traffic to fill its buffers. Leaving a lot of medium time free. When the capacity is half (NV2) then these bursts are half and thus take twice the time and thus halves the medium time for other traffic. Etc. etc. Maybe now there is a bit less package loss but that is well compensated by the double as high amount of data that can be transported.
Off course gamers and or a lot of VIOP need other approaches and NV2 might be better. But in hour network gamers (that need the best connection) are almost none existing. And Voip? Everybody uses Whatsapp and Skype and it seems people are used to be it not the same as a 'real' phone line. I at least hardly ever get complaints and when I make a call from my fiber backbone connected router to my mum in an other country that has fiber I still see no difference then with a client on my WISP network...)
Show your appreciation of this post by giving me Karma! Thanks.

Rudy R. Puister

WISP operator based on MT routerboard & ROS.
 
aerosmith9110
newbie
Posts: 32
Joined: Fri Oct 05, 2018 3:39 am

Re: ARM devices and NV2 protocol

Fri Jan 18, 2019 12:57 am

I think these could easily been avoided all by indicated in the spec sheet:

Model:
xxxxxxx
Transfer speed via wireless link ( ideal condition )
802 - 400mbps both directions TCP
nv2 -100mbps both directions TCP
nsextreme - 120mbps both directions TCP

something like that. Then the customers would have an informed decision basing on the information given and there will be no threads like this as it would simply be dismissed as " Slower speeds are due to the protocol overhead or it was already indicated in the spec sheets "
Doing this would be a vendor promise which could not be guarateed.

Thus, the disclaimer " ideal condition "
 
mfr476
Member Candidate
Member Candidate
Posts: 116
Joined: Thu Oct 11, 2018 4:51 pm

Re: ARM devices and NV2 protocol

Fri Jan 18, 2019 11:16 am

New firmware any wireless referente... Any Tri It with wireless?
 
User avatar
honzam
Forum Guru
Forum Guru
Posts: 2276
Joined: Wed Feb 27, 2008 10:27 pm
Location: Czech Republic

Re: ARM devices and NV2 protocol

Fri Jan 18, 2019 11:45 am

New firmware any wireless referente... Any Tri It with wireless?
Please use google translate
LAN, FTTx, Wireless. ISP operator
 
mfr476
Member Candidate
Member Candidate
Posts: 116
Joined: Thu Oct 11, 2018 4:51 pm

Re: ARM devices and NV2 protocol

Fri Jan 18, 2019 11:55 am

new firmware no reference to wireless. Has anyone tried if the arm improves?
 
n21roadie
Forum Guru
Forum Guru
Posts: 1886
Joined: Fri Aug 07, 2009 10:36 pm
Location: Limerick,Ireland

Re: ARM devices and NV2 protocol

Fri Jan 18, 2019 4:03 pm

@WirelessRudy
...When running csma with properly configure rts/cts I see for traffic to only 2 or 3 clients hardly any package losses. The new csma+rts/cts protocol in 'ac' works pretty fine.......
(1) Are you using just 802.11 and not Nstreme,
(2) Properly configured rts/cts - any examples you can give,
(3) new csma+rts/cts protocol - is it working good on "N".
 
mfr476
Member Candidate
Member Candidate
Posts: 116
Joined: Thu Oct 11, 2018 4:51 pm

Re: ARM devices and NV2 protocol

Fri Jan 18, 2019 4:38 pm

how can I configure my ap in csma + rts / cts protocol?
 
WirelessRudy
Forum Guru
Forum Guru
Posts: 3074
Joined: Tue Aug 08, 2006 5:54 pm
Location: Spain

Re: ARM devices and NV2 protocol

Sun Jan 20, 2019 3:09 pm

@WirelessRudy
...When running csma with properly configure rts/cts I see for traffic to only 2 or 3 clients hardly any package losses. The new csma+rts/cts protocol in 'ac' works pretty fine.......
(1) Are you using just 802.11 and not Nstreme,
(2) Properly configured rts/cts - any examples you can give,
(3) new csma+rts/cts protocol - is it working good on "N".
1) I have every P2MP network prepared for all 3 protocols and each clients listens to the AP's frequency only. Then I try each of the protocol while running some BT test from 2-4 clients to see which give the best results. NV2 alsways looses... Most of the time 802.11 is the best but some of my AP's show nstreme slightly better. (I have protection management 'enabled' and sometimes see several clients flip because of that and then in nstreme mode its much better. So this AP goes to nstreme....)
2) rtx/cts has to be 'always on' and other wireless robustness 'on' (adaptive-noise-immunity=client-mode guard-interval=long hw-protection-mode=rts-cts hw-retries=5 hw-protection-threshold=0 preamble-mode=long )
3) For 'ac' standard the csma+rts/cts has been re-designed by the IEEE and has been improved. But I don't think that works in 'N' protocol working network. The new protocol is downwards compatible if the AP is 'ac' where the CPE is still 'legacy' 'n'. But off course this 'n' cpe cannot directly benefit from the improved 'ac' protocol. But in a pre-dominant 'ac' network with a occasional 'n' unit these last will benefit too since the rest of the network works better and thus more 'airtime' is available for the 'n' units.....

The Wiki explains exactly how the different protocols have to be set. And in Google you can find loads of information about what the improvements of 'ac' are compared to 'n'. Basically that was the reason I started to try 802.11 ac whic already gave me such good results with Mimosa. And I also tried is because NV2 usually really sucks.....
Show your appreciation of this post by giving me Karma! Thanks.

Rudy R. Puister

WISP operator based on MT routerboard & ROS.
 
mfr476
Member Candidate
Member Candidate
Posts: 116
Joined: Thu Oct 11, 2018 4:51 pm

Re: ARM devices and NV2 protocol

Mon Jan 21, 2019 10:10 am

I work better in nstream Too. I can't work in 802.11 because i use voip. The problem is poor capacity...
 
WirelessRudy
Forum Guru
Forum Guru
Posts: 3074
Joined: Tue Aug 08, 2006 5:54 pm
Location: Spain

Re: ARM devices and NV2 protocol

Mon Jan 21, 2019 6:19 pm

I work better in nstream Too. I can't work in 802.11 because i use voip. The problem is poor capacity...
"VOIP" as in real phonelines over IP? Or Skype, facetime, whatsapp, etc. etc. That's also voip and work fine over my 802.11 AP's...

But I'd presume these VOIP calls are relative rare occasions. And people probably are used to poor connections anyway, especial if they use it over their Wifi.
Real VOIP as replacement for a real fixed copper line is a bit more sensitive I'd agree.

I don't understand what you mean with
poor capacity
, as far as I can see its almost the same as 802.11 and basically double that of NV2..
Show your appreciation of this post by giving me Karma! Thanks.

Rudy R. Puister

WISP operator based on MT routerboard & ROS.
 
mfr476
Member Candidate
Member Candidate
Posts: 116
Joined: Thu Oct 11, 2018 4:51 pm

Re: ARM devices and NV2 protocol

Tue Jan 22, 2019 3:43 pm

My old N cpe mikrotik works faster than my new ac cpe.
 
bastys
just joined
Posts: 10
Joined: Sun Oct 21, 2018 2:25 pm

Re: ARM devices and NV2 protocol

Tue Jan 22, 2019 4:11 pm

Just for sure, speed is measured from CPE btest or via pc for cpe? According to my measurements in certain modes, the btest does not work well, especially for new arm processors or even a combination of arm - mips.
 
mistry7
Forum Guru
Forum Guru
Posts: 1243
Joined: Tue Oct 13, 2009 11:57 am
Location: Germany

Re: ARM devices and NV2 protocol

Tue Jan 22, 2019 5:25 pm

My old N cpe mikrotik works faster than my new ac cpe.
Mikrotik call this feature......
 
User avatar
Alessio Garavano
Member
Member
Posts: 304
Joined: Sat May 29, 2004 12:49 am
Location: Corrientes, Argentina
Contact:

Re: ARM devices and NV2 protocol

Tue Jan 22, 2019 8:40 pm

Hi, this is our scenary and proper experience...

We have a backbone with 5 towers in a 10° beam angle at around 3 to 8 km of distance from our main tower, all concentrated in a PTMP Dynadish (upgraded to Level 4 to work in AP-bridge mode)...

The setting for the best performance, througput, latency, etc is all in 5Ghz-Only-N, 20/40Mhz eC, NV2 with dynamic-downlink=75% and TDMA period size=4ms, SGI, Short Preamble, MCS12 max, Frame Life-time=4, Hw-Retries=5, Inmunity activated ("ap and client" in the dynadish, "client" in remotes sites), etc...

In peak time is around 120-140Mb of througput, stable and with good latency...

Tested in Nstreme the througput is similar, but have more packet loss and higher latency.

In 802.11 is a disaster... (yesterday i test again with the WirelessRudy recommendations and the same, nothing good)

Before the Dynadish we test a LHG XL 5ac (low-cost arm cpu) and was a disaster in NV2, Nstreme and 802.11...

Today our best MT combination is mipsbe, 802.11n and NV2...
Soon we think to change all the stations of the remote towers (SXTsq, LHG and LHG XL) to dynadish, to complete the 802.11ac mipsble solution...

Is something doing this PTMP scenary pure mipsbe in 802.11ac to share experience?

Regards,
You do not have the required permissions to view the files attached to this post.
Alessio Garavano
http://www.isparg.com.ar
 
WirelessRudy
Forum Guru
Forum Guru
Posts: 3074
Joined: Tue Aug 08, 2006 5:54 pm
Location: Spain

Re: ARM devices and NV2 protocol

Wed Jan 23, 2019 1:34 am

@Alessio Garavano; Can you explain your setup again? Don't understand how it is....

You have one main backbone AP that serves at 10° 5 Dynadish stations that in return feed a tower to feed clients?
Show your appreciation of this post by giving me Karma! Thanks.

Rudy R. Puister

WISP operator based on MT routerboard & ROS.
 
alejosalmon
just joined
Posts: 23
Joined: Sun May 31, 2015 3:02 pm

Re: ARM devices and NV2 protocol

Wed Jan 23, 2019 6:03 am

I have the same idea,but I want to use some stationbox cc xl 19 dbi with 922 card inside.The main reason is that the 90°,120° antennas make too much noise in my tower.The question,is it correct due to the very little opening angle?.
 
mistry7
Forum Guru
Forum Guru
Posts: 1243
Joined: Tue Oct 13, 2009 11:57 am
Location: Germany

Re: ARM devices and NV2 protocol

Wed Jan 23, 2019 8:46 am

I have the same idea,but I want to use some stationbox cc xl 19 dbi with 922 card inside.The main reason is that the 90°,120° antennas make too much noise in my tower.The question,is it correct due to the very little opening angle?.
Don’t buy Plastic Antennas like Dynadish, look for RF Elements Horns! Next benefit if you leave the Mikrotik train you only have to exchange the adapter + Elektronic, the Antenna could stay in place
 
server8
Member
Member
Posts: 372
Joined: Fri Apr 22, 2011 1:27 pm

Re: ARM devices and NV2 protocol

Wed Jan 23, 2019 10:51 am

alejosalmon we use QRT as AP and it works great we shield it on the tower https://www.interprojekt.it/anti-noise- ... -1751.html so we don't have any problem with colocation
 
WirelessRudy
Forum Guru
Forum Guru
Posts: 3074
Joined: Tue Aug 08, 2006 5:54 pm
Location: Spain

Re: ARM devices and NV2 protocol

Wed Jan 23, 2019 11:41 am

alejosalmon we use QRT as AP and it works great we shield it on the tower https://www.interprojekt.it/anti-noise- ... -1751.html so we don't have any problem with colocation
QRT with 10 degree coverage as AP? You must have many of these as sector or all clients close together....?
Show your appreciation of this post by giving me Karma! Thanks.

Rudy R. Puister

WISP operator based on MT routerboard & ROS.
 
WirelessRudy
Forum Guru
Forum Guru
Posts: 3074
Joined: Tue Aug 08, 2006 5:54 pm
Location: Spain

Re: ARM devices and NV2 protocol

Wed Jan 23, 2019 11:44 am

I have the same idea,but I want to use some stationbox cc xl 19 dbi with 922 card inside.The main reason is that the 90°,120° antennas make too much noise in my tower.The question,is it correct due to the very little opening angle?.
Don’t buy Plastic Antennas like Dynadish, look for RF Elements Horns! Next benefit if you leave the Mikrotik train you only have to exchange the adapter + Elektronic, the Antenna could stay in place
I never understood the purpose of Dynadish anyway. QRT is much better and cheaper. Dynadish has huge sidelobs so don't work well in colocation with other AP when frequencies are close.
Plus for QRT you can buy metal shield, always better for noise reduction then plastic dome from Dynadish.j
Show your appreciation of this post by giving me Karma! Thanks.

Rudy R. Puister

WISP operator based on MT routerboard & ROS.
 
server8
Member
Member
Posts: 372
Joined: Fri Apr 22, 2011 1:27 pm

Re: ARM devices and NV2 protocol

Wed Jan 23, 2019 11:54 am

alejosalmon we use QRT as AP and it works great we shield it on the tower https://www.interprojekt.it/anti-noise- ... -1751.html so we don't have any problem with colocation
QRT with 10 degree coverage as AP? You must have many of these as sector or all clients close together....?
Yes we have a lot of QRT on the tower you can use with an angle of 15° less interference and higher perfomance
 
WirelessRudy
Forum Guru
Forum Guru
Posts: 3074
Joined: Tue Aug 08, 2006 5:54 pm
Location: Spain

Re: ARM devices and NV2 protocol

Wed Jan 23, 2019 2:21 pm

alejosalmon we use QRT as AP and it works great we shield it on the tower https://www.interprojekt.it/anti-noise- ... -1751.html so we don't have any problem with colocation
QRT with 10 degree coverage as AP? You must have many of these as sector or all clients close together....?
Yes we have a lot of QRT on the tower you can use with an angle of 15° less interference and higher performance
Ok, we have one tower serving some 100+ clients with 8 sectors of different range. All RF-Elements apart from one that we have with a 10 degree 23dBi Mars antena with rb922.
All other AP's are Netmetals on RF-Elements. We have 3x30, 2x40, 2x60 degree and one 10 degree 23dBi Mars antena with rb922.
Some AP's have still a mix of 'n' and 'ac' protocol clients. SXT's or LHG's and some Sextants.
Some other AP's are already converted in 'full' 'AC' network with SXT-sq's (arm), SXT-ac (mipsbe), DISC (arm) and LHG-5ac (arm)
All working in 5Ghz band with 40Mhz channel width.

Same tower has 7 backhauls running from it all in the 5Ghz band (1x ubnt Airfibre5x) and in 20 to 50Mhz wide bands.

At 6 meters distance we have tower of competition with 3 x ubnt sector and 1 x Airfibre5 and one ubnt ac backhaul.
At 150 meters distance we have tower of other competition with 4 x ubnt sector and 2 x ubnt backhaul.

Plenty of noise in that area....

All my AP's working in either N/AC mode or full AC mode in 802.11 rts/cts mode.
Clients can get 50Mbps and last night I ran that speed from 3 clients on the same sector both in 'n' or 'ac' and there was almost a sustained 50Mbps to these clients. I also ran a ping to one of these clients and with 30ms average wity a max of 100ms and only an occasional package drop I think it runs pretty good.

I cannot see any difference between 'arm' or 'mipsbe' devices.
Any of the AP's that is set to NV2 gets half the capacity and half the speed to the client. I tried all possible configuration options to no avail, nstreme and 802.11 outperform NV2 by 100%

If I do a test between CPE and a CCR router behind the netmetal most client have 100+Mbsp of airspeed available. Two clients at the same time still see both getting some 60 or 70. With 3 it drops to around 50 and with 4 it falls back to 20-30 each. More then 2 simultaneous sees the station with the poorest conn. rate drop its speed first.
All signals are in the -35 to -60 range with an occasional -65. (My aim is -55 or better but try not stronger then -40 and worse then -60 is actually up for antena change (if possible).)
Show your appreciation of this post by giving me Karma! Thanks.

Rudy R. Puister

WISP operator based on MT routerboard & ROS.
 
mfr476
Member Candidate
Member Candidate
Posts: 116
Joined: Thu Oct 11, 2018 4:51 pm

Re: ARM devices and NV2 protocol

Wed Jan 23, 2019 3:38 pm

share configuration please
 
WirelessRudy
Forum Guru
Forum Guru
Posts: 3074
Joined: Tue Aug 08, 2006 5:54 pm
Location: Spain

Re: ARM devices and NV2 protocol

Wed Jan 23, 2019 4:06 pm

share configuration please
who?
Show your appreciation of this post by giving me Karma! Thanks.

Rudy R. Puister

WISP operator based on MT routerboard & ROS.
 
alejosalmon
just joined
Posts: 23
Joined: Sun May 31, 2015 3:02 pm

Re: ARM devices and NV2 protocol

Wed Jan 23, 2019 7:14 pm

WirelessRudy you say that you can't see any difference between arm devices and mipsbe ones,but there is a lots complaints about packet loss in that architecture by using 802.11,nv2,nstreme.Please tell us more about it.Thanks
 
WirelessRudy
Forum Guru
Forum Guru
Posts: 3074
Joined: Tue Aug 08, 2006 5:54 pm
Location: Spain

Re: ARM devices and NV2 protocol

Wed Jan 23, 2019 8:02 pm

WirelessRudy you say that you can't see any difference between arm devices and mipsbe ones,but there is a lots complaints about packet loss in that architecture by using 802.11,nv2,nstreme.Please tell us more about it.Thanks
Exactly! Because I don't see that difference I did mention it....

Just for the sake, I ran 2 bw tests; one initiated (download, 1 tcp stream) from a rb911Lite5 ac (mipsbe) and one same parameters from DISC Lit5 ac
Ap has 14 units associated and works in 20Ce/ac protocol (All associated units are 'ac')
Bandwith test runs through 2 wireless backhauls and 2 CCR's for mpls routing towards gateway (CCR) which has a queue for each of this client set to 50Mb burst and 33 regular (we sell 30Mb)
It is also near prime time and both backhauls transport traffic of some 300 clients....

From both I run the bw test.
The rb911 has a connection rate (download) of some 270Mbps with a CCQ regurlary touching 100%. Signal is -23 (way too much but ok, It at 150 meters distance, 30 meters below.)
The DISC has a connection rate (download) of some 243Mbps with a CCQ 60-85%. Signal is -54. (This units 'looks' through a tower with 5Ghz ubnt sectors and 2 backhauls at 250 meters and the client is a 3km)

The rb911 runs a relative stable 28Mbps download (25 mins)
The DISC runs a les stable but average 23Mbps download (same time)
The AP shows other traffic from other 12 clients and is continiously showng 50-80Mbps

Ping with 256 package size and 500ms timeout:
DISC shows 25ms average with 2% package loss out of 1000 packages now
RB911 shows 13ms average with 1% package loss out of 1000 packages...

The small difference at the expense of the DISC is due its weaker signal and lower CCQ (The signal path is 15 times longer and the noise must be horrendous going to shoot 'through' another 4 sector ubnt tower. I actually picked the worse client from this sector I now realize...)

So I think the DISC Lite5ac is working absolutely fine in this 802.11ac environment. (I had the bw test running for almost 30 minutes while writing this post.)

DISC
/interface wireless
set [ find default-name=wlan1 ] adaptive-noise-immunity=client-mode allow-sharedkey=no ampdu-priorities=0 amsdu-limit=8192 \
    amsdu-threshold=8192 antenna-gain=0 area="" arp=enabled arp-timeout=auto band=5ghz-n/ac basic-rates-a/g=6Mbps bridge-mode=enabled \
    channel-width=20/40/80mhz-XXXX compression=no country=no_country_set default-ap-tx-limit=0 default-authentication=no \
    default-client-tx-limit=0 default-forwarding=yes disable-running-check=no disabled=no disconnect-timeout=3s distance=dynamic \
    frame-lifetime=0 frequency=5180 frequency-mode=superchannel frequency-offset=0 guard-interval=any hide-ssid=no ht-basic-mcs=\
    mcs-0,mcs-1,mcs-2,mcs-3,mcs-4,mcs-5,mcs-6,mcs-7 ht-supported-mcs="mcs-0,mcs-1,mcs-2,mcs-3,mcs-4,mcs-5,mcs-6,mcs-7,mcs-8,mcs-9,mcs-\
    10,mcs-11,mcs-12,mcs-13,mcs-14,mcs-15,mcs-16,mcs-17,mcs-18,mcs-19,mcs-20,mcs-21,mcs-22,mcs-23" hw-fragmentation-threshold=\
    disabled hw-protection-mode=rts-cts hw-protection-threshold=0 hw-retries=5 installation=any interworking-profile=disabled \
    keepalive-frames=enabled l2mtu=1600 mac-address=B8:69:F4:89:1A:2E max-station-count=2007 mode=station mtu=1500 \
    multicast-buffering=enabled multicast-helper=default name=wlan1 nv2-cell-radius=30 nv2-downlink-ratio=50 nv2-mode=\
    dynamic-downlink nv2-preshared-key=erwtennet@MC nv2-qos=default nv2-queue-count=2 nv2-security=enabled nv2-sync-secret="" \
    on-fail-retry-time=100ms preamble-mode=long radio-name=C4e-109 rate-selection=advanced rate-set=default rx-chains=0,1 scan-list=\
    5615 secondary-channel="" security-profile=default ssid="" station-bridge-clone-mac=00:00:00:00:00:00 station-roaming=enabled \
    supported-rates-a/g=6Mbps,9Mbps,12Mbps,18Mbps,24Mbps,36Mbps,48Mbps,54Mbps tdma-period-size=2 tx-chains=0,1 tx-power-mode=default \
    update-stats-interval=disabled vht-basic-mcs=mcs0-7 vht-supported-mcs=mcs0-9,mcs0-9,mcs0-9 vlan-id=1 vlan-mode=no-tag \
    wds-cost-range=50-150 wds-default-bridge=none wds-default-cost=100 wds-ignore-ssid=no wds-mode=disabled wireless-protocol=any \
    wmm-support=enabled wps-mode=push-button
AP
/interface wireless
set [ find default-name=wlan1 ] adaptive-noise-immunity=ap-and-client-mode allow-sharedkey=no ampdu-priorities=0 amsdu-limit=8192 amsdu-threshold=8192 antenna-gain=0 area="" arp=enabled \
    arp-timeout=auto band=5ghz-onlyac basic-rates-a/g="" bridge-mode=enabled channel-width=20/40mhz-Ce compression=no country=no_country_set default-ap-tx-limit=0 default-authentication=no \
    default-client-tx-limit=0 default-forwarding=no disable-running-check=no disabled=no disconnect-timeout=3s distance=dynamic frame-lifetime=0 frequency=5615 frequency-mode=superchannel \
    frequency-offset=0 guard-interval=long hide-ssid=no ht-basic-mcs=mcs-6 ht-supported-mcs=mcs-6,mcs-11,mcs-12,mcs-13 hw-fragmentation-threshold=disabled hw-protection-mode=cts-to-self \
    hw-protection-threshold=0 hw-retries=7 installation=any interworking-profile=disabled keepalive-frames=enabled l2mtu=1600 mac-address=CC:2D:E0:FF:FF:68 max-station-count=2007 mode=ap-bridge \
    mtu=1500 multicast-buffering=enabled multicast-helper=default name=wlan1 nv2-cell-radius=10 nv2-downlink-ratio=80 nv2-mode=dynamic-downlink nv2-preshared-key=erwtennet@MC nv2-qos=default \
    nv2-queue-count=2 nv2-security=enabled nv2-sync-secret="" on-fail-retry-time=100ms preamble-mode=long radio-name=C4-AP5 rate-selection=advanced rate-set=configured rx-chains=0,1 scan-list=\
    5550-5700 secondary-channel="" security-profile=802.11 ssid=MaruCom-C4e station-bridge-clone-mac=00:00:00:00:00:00 station-roaming=enabled supported-rates-a/g=6Mbps tdma-period-size=auto \
    tx-chains=0,1 tx-power-mode=default update-stats-interval=disabled vht-basic-mcs=mcs0-7 vht-supported-mcs=mcs0-7,mcs0-7,mcs0-7 vlan-id=1 vlan-mode=no-tag wds-cost-range=50-150 \
    wds-default-bridge=none wds-default-cost=100 wds-ignore-ssid=no wds-mode=disabled wireless-protocol=802.11 wmm-support=enabled wps-mode=push-butto
n
Show your appreciation of this post by giving me Karma! Thanks.

Rudy R. Puister

WISP operator based on MT routerboard & ROS.
 
alejosalmon
just joined
Posts: 23
Joined: Sun May 31, 2015 3:02 pm

Re: ARM devices and NV2 protocol

Wed Jan 23, 2019 10:09 pm

Perfect.What version of routeros are you using in arm and mispbe cpes? Thanks for sharing with us your results.
 
djvolt
newbie
Posts: 44
Joined: Mon Dec 31, 2018 4:50 pm

Re: ARM devices and NV2 protocol

Wed Jan 23, 2019 10:47 pm

We have dishes alu for LDF AC, LHG AC and others.

240Mbps stable on 80MHz? How many we can reach at 40MHz? 50Mbps? :P
http://24.multimediahd.pl/pl/3-60ghz
Upgrades for Mikrotik's devices
 
WirelessRudy
Forum Guru
Forum Guru
Posts: 3074
Joined: Tue Aug 08, 2006 5:54 pm
Location: Spain

Re: ARM devices and NV2 protocol

Wed Jan 23, 2019 11:24 pm

Perfect.What version of routeros are you using in arm and mispbe cpes? Thanks for sharing with us your results.
6.43.8
Show your appreciation of this post by giving me Karma! Thanks.

Rudy R. Puister

WISP operator based on MT routerboard & ROS.
 
WirelessRudy
Forum Guru
Forum Guru
Posts: 3074
Joined: Tue Aug 08, 2006 5:54 pm
Location: Spain

Re: ARM devices and NV2 protocol

Wed Jan 23, 2019 11:25 pm

We have dishes alu for LDF AC, LHG AC and others.

240Mbps stable on 80MHz? How many we can reach at 40MHz? 50Mbps? :P
If you get 240Mbps on 80MHz then you get half of that on 40Mhz....
Show your appreciation of this post by giving me Karma! Thanks.

Rudy R. Puister

WISP operator based on MT routerboard & ROS.
 
djvolt
newbie
Posts: 44
Joined: Mon Dec 31, 2018 4:50 pm

Re: ARM devices and NV2 protocol

Wed Jan 23, 2019 11:32 pm

Really? No way man, Ubnt AC 240-280Mbps on 40Mhz at 8km :P Around 130Mbps at 20MHz...
http://24.multimediahd.pl/pl/3-60ghz
Upgrades for Mikrotik's devices
 
WirelessRudy
Forum Guru
Forum Guru
Posts: 3074
Joined: Tue Aug 08, 2006 5:54 pm
Location: Spain

Re: ARM devices and NV2 protocol

Thu Jan 24, 2019 11:04 am

Really? No way man, Ubnt AC 240-280Mbps on 40Mhz at 8km :P Around 130Mbps at 20MHz...
I beleive 130 is roughly half of 240-280........ :o
Show your appreciation of this post by giving me Karma! Thanks.

Rudy R. Puister

WISP operator based on MT routerboard & ROS.
 
mistry7
Forum Guru
Forum Guru
Posts: 1243
Joined: Tue Oct 13, 2009 11:57 am
Location: Germany

Re: ARM devices and NV2 protocol

Thu Jan 24, 2019 11:04 am

Really? No way man, Ubnt AC 240-280Mbps on 40Mhz at 8km :P Around 130Mbps at 20MHz...
Mimosa too, but Mikrotik maxes out with 180-190 Mbit @ 40 Mhz
 
WirelessRudy
Forum Guru
Forum Guru
Posts: 3074
Joined: Tue Aug 08, 2006 5:54 pm
Location: Spain

Re: ARM devices and NV2 protocol

Thu Jan 24, 2019 11:21 am

Really? No way man, Ubnt AC 240-280Mbps on 40Mhz at 8km :P Around 130Mbps at 20MHz...
Mimosa too, but Mikrotik maxes out with 180-190 Mbit @ 40 Mhz
speeds are a factor of bandwidth, the amount of clients, protocol used, signals strength (=MCS rate) and package losses (=interference)

In lab without interference, good signals and only one client you get roughly 50% of the conn. rate as throughput speed. Minus overhead due protocol and other stuff 45-55% tcp is more realistic speed compared to the conn. rate.
The conn. rates more then doubles by doubling the channel bandwidth. Resulting in almost doubling the available tcp throughput.
If you get 100Mbps with 20Mhz then you get 200Mbps with 40, 400Mbps with 80Mhz etc.

Mikrotik only has one 750Mhz cpu where Mimosa works with 1,2Ghz I believe plus they have a better firmware. Also their antennas are better then the usually Mikrotik setups...
I would say Mikrotik theoretically can do some 70-80% of what Mimosa can do. But the more clients per AP the faster Mikrotik looses capacity compared to Mimosa.
(These are personal opinions. I have 4x Mimosa A5 with some 80 clients in a dense region working in 80Mhz band and in a wider area some 600 Mikrotik clients and some 40 AP's.
I also have one eCambium2000 setup with some 8 clients, some SXT elevate, some ubnt elevate and some eCambium stuff.. In other words, I can compare things a bit. Al tough not a situation is the same.)
Show your appreciation of this post by giving me Karma! Thanks.

Rudy R. Puister

WISP operator based on MT routerboard & ROS.
 
marcin21
Member Candidate
Member Candidate
Posts: 185
Joined: Tue May 04, 2010 4:50 pm

Re: ARM devices and NV2 protocol

Thu Jan 24, 2019 4:01 pm

@WirelessRudy, how would You compare effectiveness of Elevate vs original ROS (% of bndwdth/jitter) ?
And how would You compare 80mhz ptmp A5c vs mikrotik AC?
As my networki is rather old, I could say I'm building it since beginnig of Mtik adventure in this field, I've got a lot towers mikrotik only. Trying to keep it up to date, swapped
AP to OmnitiksAC that gave me rather headache than improvement so I wonder what steps should I take in face of nstreme/nv2 "development" process.
And Your experience in this area seems to be very interesting for me, so any answer would be very welcome. :)
Best regards
 
mfr476
Member Candidate
Member Candidate
Posts: 116
Joined: Thu Oct 11, 2018 4:51 pm

Re: ARM devices and NV2 protocol

Thu Jan 24, 2019 5:24 pm

Ac is "mikrotik 10 year challenge"
 
WirelessRudy
Forum Guru
Forum Guru
Posts: 3074
Joined: Tue Aug 08, 2006 5:54 pm
Location: Spain

Re: ARM devices and NV2 protocol

Thu Jan 24, 2019 6:49 pm

@WirelessRudy, how would You compare effectiveness of Elevate vs original ROS (% of bndwdth/jitter) ?
And how would You compare 80mhz ptmp A5c vs mikrotik AC?
As my networki is rather old, I could say I'm building it since beginnig of Mtik adventure in this field, I've got a lot towers mikrotik only. Trying to keep it up to date, swapped
AP to OmnitiksAC that gave me rather headache than improvement so I wonder what steps should I take in face of nstreme/nv2 "development" process.
And Your experience in this area seems to be very interesting for me, so any answer would be very welcome. :)
Best regards
My network was 100% and is now some 80% based on Mikrotik. I started with Mikrotik in 2004 I believe. But since about 6 years ago started to look around and decided to try Mimosa in our neigbourhood that now has fiber as well. Only because I swapped to Mimosa and offer 50, 100 and 200Mbps package to my clients here for a little bit less then the fiber company charges I still have all my clients here. When we would have stayed with the crappy Mikroik NV2 from 2/3 years ago I'd probably had lost all my clients to fiber.....

Now it seems with the new line of 'ac' models with faster cpu and (slow-) improvements in ROS and I'd massively updating several of my P2MP networks into 'ac' I can start to offer 30 and 50Mbps in the countryside.

When it comes to "comparing" is my experiences and looking at the results between the different brands and way it's used.
Mimosa was from day one (3 years ago) 'ac' and their tdma is definately better then that of Mikrotik.

We also have one eCambium2000 with beamforming antenna (both 802.11'n') that we won in a raffle and we use with some 8 clients. Although I wouldn't have bought it for the investment it works good. Clients are just connected and have good speeds.
(So why not change into eCambium? Well, I have no money tree in my backgarden. Changing a P2MP network from one vendor into another is not only changing an AP, but also all clients... meaning a heavy extra investment that also needs a lot of time (only one installer) and during the swap you need to make sure clients still have a connection....)
Show your appreciation of this post by giving me Karma! Thanks.

Rudy R. Puister

WISP operator based on MT routerboard & ROS.
 
marcin21
Member Candidate
Member Candidate
Posts: 185
Joined: Tue May 04, 2010 4:50 pm

Re: ARM devices and NV2 protocol

Fri Jan 25, 2019 2:14 am

@WirelessRudy
thank You for reply :)
Do I got You right, that on 80mhz ptmp scenario You can supply clients with 200mbps?
ofcourse I do understand it's real world so overbooking is taken into account. But no complains from customers is enough , isn't it?
 
WirelessRudy
Forum Guru
Forum Guru
Posts: 3074
Joined: Tue Aug 08, 2006 5:54 pm
Location: Spain

Re: ARM devices and NV2 protocol

Fri Jan 25, 2019 11:21 am

@WirelessRudy
thank You for reply :)
Do I got You right, that on 80mhz ptmp scenario You can supply clients with 200mbps?
ofcourse I do understand it's real world so overbooking is taken into account. But no complains from customers is enough , isn't it?
With Mimosa I have no issue in reaching 200mbps for my clients. In Mikrotik I don't know. I only did a rough test the other day where the 80mhz band overlapped at least two other sectors on the same tower and could run 175Mbps at peaks from a LHG-5ac but maybe with more careful picking of a frequency and finetuning the network I could. But it's too much time consuming to make a perfect test environment.
Show your appreciation of this post by giving me Karma! Thanks.

Rudy R. Puister

WISP operator based on MT routerboard & ROS.
 
marcin21
Member Candidate
Member Candidate
Posts: 185
Joined: Tue May 04, 2010 4:50 pm

Re: ARM devices and NV2 protocol

Fri Jan 25, 2019 12:06 pm

Yes, I'm aware that You spoke about gears with handwritten lowercase "m" on cover.
Thanks again.
Live long and prosper. :)
 
User avatar
honzam
Forum Guru
Forum Guru
Posts: 2276
Joined: Wed Feb 27, 2008 10:27 pm
Location: Czech Republic

Re: ARM devices and NV2 protocol

Tue Feb 05, 2019 3:51 pm

From support:

Hello,
Unfortunately, our development team has been unsuccessful to fix the issue with ARM ac chips combined with nv2 yet. It is still being worked on, but perhaps a new TDMA protocol will be developed to support our new chip-sets, if the issue will not be resolved. Apologies for the inconvenience.
LAN, FTTx, Wireless. ISP operator
 
mistry7
Forum Guru
Forum Guru
Posts: 1243
Joined: Tue Oct 13, 2009 11:57 am
Location: Germany

Re: ARM devices and NV2 protocol

Tue Feb 05, 2019 4:47 pm

From support:

Hello,
Unfortunately, our development team has been unsuccessful to fix the issue with ARM ac chips combined with nv2 yet. It is still being worked on, but perhaps a new TDMA protocol will be developed to support our new chip-sets, if the issue will not be resolved. Apologies for the inconvenience.
With the same Timeline as ROS 7?
lol......
 
mfr476
Member Candidate
Member Candidate
Posts: 116
Joined: Thu Oct 11, 2018 4:51 pm

Re: ARM devices and NV2 protocol

Tue Feb 05, 2019 8:17 pm

I changed all my mikrotik arm equipment for other brand and now I work fine. Arm problem would be solve in 2020.
 
okoun
Frequent Visitor
Frequent Visitor
Posts: 98
Joined: Fri Feb 18, 2011 3:11 pm

Re: ARM devices and NV2 protocol

Tue Feb 05, 2019 9:25 pm

It is a disgrace that in recent years Mikrotik on wireless cough, now is not done a spectral scan for AC !!!!
Shame
Still showing devices with ARM very poor transmission parameters with NV2, why long ago does not work on NV3 ????
Instead, new and new features are being made, but old things are coughing.
I'm crying.
 
mistry7
Forum Guru
Forum Guru
Posts: 1243
Joined: Tue Oct 13, 2009 11:57 am
Location: Germany

Re: ARM devices and NV2 protocol

Tue Feb 05, 2019 9:36 pm

It is a disgrace that in recent years Mikrotik on wireless cough, now is not done a spectral scan for AC !!!!
Shame
Still showing devices with ARM very poor transmission parameters with NV2, why long ago does not work on NV3 ????
Instead, new and new features are being made, but old things are coughing.
I'm crying.
There is enough working from other vendors on the market, buy that
 
InoX
Forum Guru
Forum Guru
Posts: 1961
Joined: Tue Jan 09, 2007 6:44 pm

Re: ARM devices and NV2 protocol

Wed Feb 06, 2019 12:39 am

A sector with 15 clients a/n/40mhz and the one I've tested is old single core .ac.
Under each picture you can find the protocol used.
Sadly in 802.11 is not stable and clients gets a lot of disconnections so I'm sticking with nstreme.
The conclusion is Mikrotik have lost speed in any protocol, a lot.
Just for fun.
You do not have the required permissions to view the files attached to this post.
 
peson
Trainer
Trainer
Posts: 178
Joined: Tue Jul 20, 2004 10:33 am
Location: Sweden

Re: ARM devices and NV2 protocol

Fri Feb 08, 2019 11:39 pm

A sector with 15 clients a/n/40mhz and the one I've tested is old single core .ac.
Under each picture you can find the protocol used.
Sadly in 802.11 is not stable and clients gets a lot of disconnections so I'm sticking with nstreme.
The conclusion is Mikrotik have lost speed in any protocol, a lot.
Just for fun.
Are the pictures showing your bandwidth tests on 802.11, NV2 and Nstrem in that order?
Reboot is the last resort, try to find out what's wrong instead.

Who is online

Users browsing this forum: No registered users and 6 guests