Just some personal considerations and one question...
First of all: I like mikrotik very much. However, I must note that it is necessary to look ahead to overcome current mikrotik limits in wireless, particularly in PtMP.
I mean: fixed line's competitors are leading to customers' house 30 or 50 Mbit, while the 5GHz spectrum is very dirty today. With current technology, in real environment, we can bring up to 20Mbit for a single AP when we're lucky.

Mikrotik seems doesn't offer any guarantees in wireless development...it is nearly blocked: no idea when we can have good filters, MU-MIMO, sync, better hardware, more bandwidth, etc...

A question to the Mikrotik's staff: do you believe in WISP's market yet?

I don't think the issue is Mikrotik.

I think the issue is the regulatory guys like the FCC and the current wireless technology available to the public.

Also the fact that most of the world is rolling out fiber services to the home and business can be the cause for the wireless development slowing down.

I think all the big players are focused on pushing more data over fiber and indoor wireless at the moment and they do not see the need to work on 5Ghz wireless equipment at the moment.

Hopefully someone makes a discovery soon that could give new life to 5Ghz

You're right, but I don't think WISP market is dead...
I've tested airprism ubiquiti in lab and I must say its filter works very well. With mikrotik stuff, interference from adjacent channels and from far channels limits real bandwidth drastically.
I don't want use ubiquiti in my network, but at least they are moving forward.

For the price of the current Mikrotik's routerboard lineup it is hardly to expect more.
(802.11 a/b/g/n/ac standard by default simply isn't optimized for long outdoor p2mp configuration).
In fact, Mikrotik is very good (in its price category) but we always want more

There are some (technically speaking) better solutions from some other companies but the price of their equipment is also much much higher and they are using their own proprietary wlan standards.
We are aslo waiting for something revolutionary better for p2mp.... but not at the 10x price.

Regards,
M.

I would like wireless equipment with ubnt's fancy spectral scan and ac performance with MT Ros for routing ect.

Just my 2 cents worth, having to pay R800 south african rands that is roughly $52 plus minus, i have setup networks with over 30 links streaming 1080p video footage from over 10 cameras or more and mostly being ptmp with the exception of using some rbsxt5acsa thats a little more costly and they all work flawlessly. I never have video drops or link downtimes.

All that it boils down to is finding the most unused frequencies in that area and sticking with it. Where i live its like a massive basin and its highly saturated with 5ghz but will tweaking you can always make it work.

This here is where mikrotik falls into tho, if you like the simplicity of ubnt, then go for it. If you want the security and modding with sophisticated VLANS, routing protocols and much more, then go for Mikrotik.

Or if you are realy fussy about spectral scans that look nice, go fork out $788 for Radwin or some enterprise class equipment. IMO Mikrotik has been more stable in my line of work than ubnt products. As a bonus, you even get free dyndns with Mikrotik.

AC works awesomely btw just have to find the right frequencies and it should work. What distances are you looking at?

Hi flyingsheep,
with 10 cameras and 30 links in clean spectrum you can use all technology you want.
I'm talking about ten thousand customers and one thousand mikrotik devices in a dirty spectrum...

This here is where mikrotik falls into tho, if you like the simplicity of ubnt, then go for it. If you want the security and modding with sophisticated VLANS, routing protocols and much more, then go for Mikrotik.

Or combine both. There is not only simplicity in the UBNT AC gear. They introduced RF-Filtering, Pimped Atheros HW with an additional Chip to help with TDMA handling, added a second wireless card for background RF-scanning and they will do GPS-Sync. Software is a weak point there but the HW is a generation in front of what MT does with .ac.

we have thousands of MT radios and until we started trying to offer 16-25 meg packages all was fine. Some AP's will do 40-70 meg and others with the exact same settings will not. Not sure what that is all about. What we need is GPS sync for freq reuse and timing. I keep hoping MT will add GPS to the lineup, but thus far nothing. All our new deployments are ePMP and while i do not like the interface, the stuff works. We do ABAB channel reuse on our towers and there is never a problem. The filters are incredible. I really do not want to forklift my entire network, but will if that's what it takes to compete.
Mikrotik if you are reading, hire the talent if necessary to reverse engineer the ePMP gear. Price of AP''s is a lot more but the SM's are cheaper in some cases. I have faith in you, get it done. I love Mikrotik, you enabled me to grow to a very large and profitable company, but you need to grow with all of us.

See you at MUM in USA, hope you have some news!

Some AP's will do 40-70 meg and others with the exact same settings will not.

Band? N or ac?
Dual pol?
Channel? 20? 40? 80?
Number of clients per AP?
Signlas?


In 20mhz, dual pol N, about 20-30 clients connected, signals between -68 and -50 I can achieve max 25mbit where interferences are not strong.

When interferences in near channels are strong, about 10 max 15Mbit.
<10Mbit if they are in channel.

Band? N or ac?

We never use AC, it is fine for backhauls but worthless in a highly congested area like ours.

Dual pol?

Everything is dual polarity and most are 90 degree sectors.

Channel? 20? 40? 80?

20 MHz channel spacing in most all cases. That should allow for 100 megs of traffic and that is all i intend to push over a single AP. 40 MHz in almost impossible due to congestion. We have a couple of backhauls using 80 MHz, but in actuality i see almost the same performance over 40 MHz channel.

Number of clients per AP?

We limit AP's to a /27. Rather than adding more IP's, we add another AP.

Signlas?

Minimum signal allowed is -69. In order for N to modulate signal must be stronger than this. We shoot for -60. I wish there was an auto power down on clients so the close in SM's would back off the power. This is another feature i like of the ePMP. When i look at a ePMP AP, all the clients have signal within 5-10 db. That is the way it should be.

Frequency?

We use 2.4 GHz, 5 GHz, and 900 MHz sparingly. The majority of our deployments are 5 GHz.

On a loaded AP i have some pushing 35-40 meg, while others seeming to top out at 25-27 meg. i have found the Netmetals have a much better chance of the high throughput than say a RB912. I assume it is the processor and ram.

We have a couple of UBNT towers and on a 20 MHz channel with a good quality connections, i can pull 95 Meg to an SM. So apparently they have figured it out. And this is the old Rocket M5's. We have abandoned any new UBNT deployments, just sick and tired of all the beta software and promises that never happen. Plus I refuse to deploy a 4 channel use system. (4 sectors on 4 different freq)