Mikrotik, there are any news about improving NV2 protocol?

Today we can't have more than 25Mbps aggregate on an 20Mhz AP with 30 stations
(5Ghz n/ac)

You could implement TDD with a ratio of DL/UL.
The most widely used in networks PTMP is 75/25.

All combined with a GPS Sync or Internatl Sync dramatically improve aggregate bandwidth (80-90Mbps) and make reliable connect 70-80 stations in real environment.

Are you working on it? Any of your engineer is studying something? Do you have changes in the air? Give us some more news... your wireless market for wisp is dying

I don't believe that nv2 will be improved, nothing has improved from a long time ago. I limit the number of stations to ~20 because of high ping and small speed.

I hope normis,mrz,uldis, or someone of Mikrotik team answer to give informations on the situation.

And I hope they are good news.

We are constantly growing and limit customers to 20 per AP means losing customers

+1 for nv3

NV 3... Please...

Sent from my Che2-L11

It would be pretty cool, NV3 !!
Today's customers are more and more, few frequencies, the spectrum is more dirty, and customer bandwidth trend increase (competitors increases)

Urge then, see if Mikrotik wants to develop the radio technology for WISP, or remain, leave give me the comparison, as the Dude (stationary for years)

If yes, unfortunately they must also listen to the voice of the people, our needs.

- TDD with Custom Ratio DL/UP
- GPS Sync, or another sync system in same tower
- Modulation mechanisms more robust (why slow down rate when the are inactivity? require more time to find better modulation when traffic is generated)
- Methods of checking error more detailed (for example drop radio packet)
- Better support in nLos o NLOS (is now a total disaster)
- Hardware RF filter
- Errors correction in engineering design of hardware .. sorry but even today with some rain I was burned additional NetMetal (incoming from sticker)

But if none of this is in your thoughts, then better to gear up with something else

We have the same problem
Is there some chance to improve NV2?

With NV2/40mhz can you tell me the max throughput?
An RB912/RB922ac i think can do 50-60Mbps with 20/30 nv2 clients?

-to put GPS sync means change the APs-CPEs again
-maybe some cheaper software improvement could be made?

I think seriously to buy cambium epmp1000, with gps sync etc, but the price is min 2x Lite5, and you can put 100-120 clients/AP

We have the same problem
Is there some chance to improve NV2?

With NV2/40mhz can you tell me the max throughput?
An RB912/RB922ac i think can do 50-60Mbps with 20/30 nv2 clients?

-to put GPS sync means change the APs-CPEs again
-maybe some cheaper software improvement could be made?

I think seriously to buy cambium epmp1000, with gps sync etc, but the price is min 2x Lite5, and you can put 100-120 clients/AP

Yes epmp is a system which is considerable/comparable. MT should look at it and implement some features like gps. The SMs are higher quality than sxt light. The door mechanism of SXT is not as good and we see damaged ethernet ports as it is narrow to insert ethernet cable.
Epmp has no .ac (even not announced) so you will likely replace it in the near future. 100 cpes on one wifi based ap will give bad performance regardless how they tune the mac.

Anyone knows how the performance is in real life situation on ubiquity gear? The folders looks nice but can it perform?

Anyone knows how the performance is in real life situation on ubiquity gear? The folders looks nice but can it perform?

better than MT. Sorry for that. In our our Case we have tested both. Netmetal an UBNT AC Gear.

With a normal Nanobeam M5 we get 150 Mbit throuput over a 2 km link.

for PTMP in NV2 case, we are interested to put more CPEs to an AP
And on UBI, how may clients can you put to an AP, and what is the max troughput of the AP?

NV2 / 20mhz or dual NV2 40Mhz or Airmax5

better than MT. Sorry for that. In our our Case we have tested both. Netmetal an UBNT AC Gear.

With a normal Nanobeam M5 we get 150 Mbit throuput over a 2 km link.

Ok, 2km, so your test is a PTP.
But how can you say me on PTMP ? This is the big problem...
In my experience I see
- ubnt airmax: 50-60 stations on AP with an aggregate of 35-40Mbps
- mkt nv2: 45 clients on AP with a aggregate of 30-35Mbps (personal experience in a very good scenario)
difference?
ubnt has very fantastic built-in QoS (if match with your needs)
mkt when are near the limit (also stations per AP) start lowering bandwith and spike (also if AP is with 6-7Mbps in use)

- epmp: on AP with 70 stations with an aggregate of 80Mbps
- pmp450: on AP with 150 stations aggregate of 90-100Mbps and very low latency and not frequent spikes

Mikrotik can has the potential of improving his wireless for Wisp but, as this thread say, no one single comment by Mikrotik employees

better than MT. Sorry for that. In our our Case we have tested both. Netmetal an UBNT AC Gear.

With a normal Nanobeam M5 we get 150 Mbit throuput over a 2 km link.

Ok, 2km, so your test is a PTP.
But how can you say me on PTMP ? This is the big problem...
In my experience I see
- ubnt airmax: 50-60 stations on AP with an aggregate of 35-40Mbps
- mkt nv2: 45 clients on AP with a aggregate of 30-35Mbps (personal experience in a very good scenario)
difference?
ubnt has very fantastic built-in QoS (if match with your needs)
mkt when are near the limit (also stations per AP) start lowering bandwith and spike (also if AP is with 6-7Mbps in use)

- epmp: on AP with 70 stations with an aggregate of 80Mbps
- pmp450: on AP with 150 stations aggregate of 90-100Mbps and very low latency and not frequent spikes

Mikrotik can has the potential of improving his wireless for Wisp but, as this thread say, no one single comment by Mikrotik employees

Matches with what we see. MT is good for low density. Epmp does scale much better esp. as gps allows frequency reuse. Pmp450 is best when price is no problem.

So we install MT at small sites and epmp when we expect higher numbers.

MT has still the advantage of ROS and the flexibility to put/exchange boards.

It would be pretty cool, NV3 !!
Today's customers are more and more, few frequencies, the spectrum is more dirty, and customer bandwidth trend increase (competitors increases)

Urge then, see if Mikrotik wants to develop the radio technology for WISP, or remain, leave give me the comparison, as the Dude (stationary for years)

If yes, unfortunately they must also listen to the voice of the people, our needs.

- TDD with Custom Ratio DL/UP
- GPS Sync, or another sync system in same tower
- Modulation mechanisms more robust (why slow down rate when the are inactivity? require more time to find better modulation when traffic is generated)
- Methods of checking error more detailed (for example drop radio packet)
- Better support in nLos o NLOS (is now a total disaster)
- Hardware RF filter
- Errors correction in engineering design of hardware .. sorry but even today with some rain I was burned additional NetMetal (incoming from sticker)

But if none of this is in your thoughts, then better to gear up with something else

+1

.. .. listen to the voice of the people, our needs.

- TDD with Custom Ratio DL/UP
- GPS Sync, or another sync system in same tower
- Modulation mechanisms more robust (why slow down rate when the are inactivity? require more time to find better modulation when traffic is generated)
- Methods of checking error more detailed (for example drop radio packet)
- Better support in nLos o NLOS (is now a total disaster)
- Hardware RF filter
- Errors correction in engineering design of hardware ..

+1

+1 for nv3

+1

Same experience here. Two APs. No NV2, just straight 802.11n (2.4GHz). 35 - 45 clients each, terrible jitter (5 - 200+ ms) and CCQ variation (15% - 90%). Noise floor -109 to -111 which is the best possible.

Client furious. Bad for my reputation.

Mikrotik, can we expect some hard work being put on this? Anyone? Normis?

Same experience here. Two APs. No NV2, just straight 802.11n (2.4GHz). 35 - 45 clients each, terrible jitter (5 - 200+ ms)

Why you have NV2 disabled? This topic is about NV2.
Enable nv2, set TDMA to auto. 35-45clients is too much for AP. Especialy on 2,4Ghz.
Move clients to 5Ghz with NV2

Same experience here. Two APs. No NV2, just straight 802.11n (2.4GHz). 35 - 45 clients each, terrible jitter (5 - 200+ ms)

Why you have NV2 disabled? This topic is about NV2.
Enable nv2, set TDMA to auto. 35-45clients is too much for AP. Especialy on 2,4Ghz.
Move clients to 5Ghz with NV2

Yes, topic is about NV2 that is proprietary Wireless protocol from Mikrotik.
So I think you talk about standard 802.11 and standard wireless devices.

My question is about improvments of NV2 (that someone call NV3) that can now has bad behaviours in Point-To-Multipoint.

I renew my question: Mikrotik any news about improving this technology? Are your engineers working on it?

Same situation, performanse degradation are obvious, and users complaining every day (

Mikrotik ??

Max 30 users per AP and max 25Mbit in better scenario with MiMO 2x2 in 802.11n, 20mhz,nv2. For me this is enough for today but non to for tomorrow.

The main problem is in the most of cases spectrum is very dirty, so max bandwith is reduced to 5-10Mbit per AP. So it is impossible to get stability and it is impossible to do business...


I think wireless isn't important for mikrotik today. I'm very worry about this.

I agree, Mikrotik not longer consider WISPs market.
In fact, until today, still no answer on this thread by them.

They are focusing on IPSEC
In the last releases 70% of the time appears ipsec fix
Stop, and fix/complete it for good!!

I agree, Mikrotik not longer consider WISPs market.
In fact, until today, still no answer on this thread by them.

They are focusing on IPSEC
In the last releases 70% of the time appears ipsec fix
Stop, and fix/complete it for good!!

Agreed ? Does anyone have market share numbers for WISP vs other internet provider platforms?

I think Mikrotik is opening to the market of home network.

In the last newsletter (67) we see a photo of new product (sector) but is not mentioned into it

We have problem when client with "one chain" antenna are connected to MIMO sector.
Speed are dramaticaly fall when "one chain" user start transfer even small traffic, pictures below:

Speedtest from base station to one Client (dual chain SXT) Speedtest from "one chain GRID" user to Base station, signal good (-65/-65) Speedtest form base station to earlier GOOD client drops (dual chain SXT)

Speedtest JUMP from base station to earlier client when "one chain" user stop transfering...

Your posts are not related with topic title. Please start your own Topic.

Your posts are not related with topic title. Please start your own Topic.

I think that my problem are NV2 related ?

Your posts are not related with topic title. Please start your own Topic.

I think that my problem are NV2 related ?

draguzet this is a typical problem that I've in NV2 PTMP.
It should not happen, or impact so strongly because of the nature of TDMA, but in Mikrotik anything could happen

Your posts are not related with topic title. Please start your own Topic.

I think that my problem are NV2 related ?

draguzet this is a typical problem that I've in NV2 PTMP.
It should not happen, or impact so strongly because of the nature of TDMA, but in Mikrotik anything could happen

We offer a free upgrade for customers with older cpes to increase sector capacity.

We offer a free upgrade for customers with older cpes to increase sector capacity.

This problem appear also on dual-chain device. So free upgrade not solve the problem

We offer a free upgrade for customers with older cpes to increase sector capacity.

This problem appear also on dual-chain device. So free upgrade not solve the problem

Yes. But it reduces the problem increasing the speed of the problematic cpes.

Not a good idea having dual chain AP and then connect single and dual chain CPE's ?

Not a good idea having dual chain AP and then connect single and dual chain CPE's ?

A single chain cpe eats up to double the airtime compared to a dual chain cpe. If it is a bad performer due to bad conditions it slows down the sector.

A single chain cpe eats up to double the airtime compared to a dual chain cpe. If it is a bad performer due to bad conditions it slows down the sector.

What I mean is that, just having one or two cpe with problems, to degrade the whole cell.

TDMA protocol should give safeties in these cases
There is somethings wrong, that could be fixed

So, I renew my request to Mikrotik staff: there are some scheduled works to improves NV2 ?

Not a good idea having dual chain AP and then connect single and dual chain CPE's ?

A single chain cpe eats up to double the airtime compared to a dual chain cpe. If it is a bad performer due to bad conditions it slows down the sector.

Unless I am mistaken TDMA has a predefined time slot for each client, the AP will listen to each client for this predefined time slot and move on to next client and when it encounters bad performer (client) this gets the same predefined time slot and AP moves to the next client, so the sector cannot be slowed down, slow downs are caused by either (or both) number
of registered clients or bandwidth given to each?

A single chain cpe eats up to double the airtime compared to a dual chain cpe. If it is a bad performer due to bad conditions it slows down the sector.

What I mean is that, just having one or two cpe with problems, to degrade the whole cell.

TDMA protocol should give safeties in these cases
There is somethings wrong, that could be fixed

So, I renew my request to Mikrotik staff: there are some scheduled works to improves NV2 ?

What do think occurs on the dual chain AP unused chain (antenna element) when the AP goes into receive mode and is listening to both dual and single chain CPE's?

Suddenly there is no signal on one chain when switching from last client, does the radio card use AGC (?) in a attempt to boost signal then onto next client, just as suddenly there is now signal on that chain, does the radio card this time use AGC to reduce the signal level,

If I am correct this would explain how and why the sector performance is reduced?

What do think occurs on the dual chain AP unused chain (antenna element) when the AP goes into receive mode and is listening to both dual and single chain CPE's?

Suddenly there is no signal on one chain when switching from last client, does the radio card use AGC (?) in a attempt to boost signal then onto next client, just as suddenly there is now signal on that chain, does the radio card this time use AGC to reduce the signal level,

If I am correct this would explain how and why the sector performance is reduced?

The single polarity certainly affect performance.
But I also have situations where a dual polarity antenna is not well aligned because the strong wind has move it and has a weak signal

In the case of draguzet I think that:

If the AP has a total capacity of 100Mbps
and station1 alone has 100Mbps
and station2 alone has 6Megabit

The TDMA divides to 50% the band (or the 50% of airtime):
if both are using the bandwidth I should expect:
station1 should make 50Mbps
station2 should make 6Mbps

The single polarity certainly affect performance.
But I also have situations where a dual polarity antenna is not well aligned because the strong wind has move it and has a weak signal
..........................

I also have CPE's that have their signal reduced from installation but haven't noticed any reduction
in AP NV2 performance but what really effects performance is the increase in the number of registered clients and the time this takes for NV2 to process the total number of timeslots.

You mean that (for example) :
if I have an AP with 40Mbps of bandwidth and 40 customers registrated, they will be castrated to 1Mbps because slots are equally distributed, whether they do or not traffic

+1 Need 75/25 TDD ratio.