We have a 100 meg light fibert from spectrum to our core mikrotik.

From the core we have a wireless link out into the boonies, using Ubiquity Rocket M's.

Internet -> core -> AP -> SM

Speed tests from SM to core show 20 megs

Speed tests from core to fast.com show 20 megs

Speedtests from SM to fast.com show 5 megs.

Why?

Homer W Smith
Lightlink CEO

capacity of the device watch, and the actual throughput.

Sorry what is a device watch? Please be patient with me, I am somewhat of a beginner at this.

And thanks for answering1

Homer

Sorry I meant watch the capacity of your access points and bridges. Each device you should be able to see it's capabilities on a given channel picked, width picked (20/40), power output and is it a PTP or PTMP. I like dslreports.com when running speed tests. PTP speeds can be amazing meaning the link is great. The speed between the links have to be backed by the megs you have at your disposal. Do not spread yourself thin. It's always to have more speed than not enough.

Yes thanks, your point is well understood.

The point is that the physical upstream has a certain amount of bandwidth available, not much but maybe 20 megs,
if I use a hard wire connection to the core router. But over the wireless link to the core router only about 5 megs is
available.

Thus those doing speed tests from the wireless network see less available than do the hard wire
connections, both connecting to the same core router before going to the internet.

The radios on the wireless are only running at about 50 percent capacity, so some efficiency is
being lost in the transition from the wireless link to the core router: wireless -> hard wire -> core -> internet link.

Homer

Hi
you can try to do speed test in the client wireless cpe to your tower ap to check your realtime throughput
whats your tx/rx rate showing in client side wireless cpe? signal strength? try to fine tune to get a gud signal
by the way rocket m5 i guess u are using is not advised for backhaul radio to get a realtime throughput of 100mbps from your upstream provider
go for powerbeam ac or mikrotik dynadish ac or netmal5ac with 30db dish...

Homer: What are your expectations of the Rocket M's? How much "speed" should they provide. Do NOT copy the marketing bla...bla... Like : "300 Mbps !!! YEAH". That is NOT realistic.

Also, they show this for ONE user. I assume you will have much mire users. So what happens if only one person wants the cake, and what happens if 10 people want the same cake. How much will each person get from the cake? Is that more or less, that if you only ONE person get's the cake?


And more technical: What frequency, how is the noise, how many overlapping channels etc...

In response to the actual capabilities you are seeing depends on what type of equipment you are utilizing. There is a big difference between a M Series Rocket to a AC Series Rocket. The M's you are looking at 75 megs up and down. That is a near perfect link. A problem is jump down to a 20 mhz for less overlap costs capacity, switch to 40 mhz is hard in a noisy environments. The AC's are expensive, and mixing the M's and AC's cut down on the performance greatly. Do you use M series or AC? I am not knocking the M series they work, and have for many years. Every area has challenges and demands.

What kind of bandwidth do you have to work with? AC units will help if you can supply enough bandwidth to back them and justify the cost. Do you do ptp or ptmp on the rockets? I have seen numbers off just one of my rockets close to 700 megs ptp. They do amazing. When I run in to frequency overlap issues in a high noise environment be it in the city, housing communities, and competition I see it almost immediately. I can monitor my system by cell phone no matter where I am. I started with with high end tech and I am barely utilizing there capabilities. That can change overnight, and it is possible I may need one link mainly diverted to a big building with many units. I have two of the airfiber 5's if needed,and 4 Rocket AC's in use. I do have M's out there too and perform great too. If I want to expand further out I know I have that capability. If I can make a customer happier, and compete with the cable companies at the same time then the customer and I win.

The cake is only so big so if a person wants a bigger piece they need to pay for it. Traffic shaping gives them what they expect and service they expect. Eventually as more customers get on buy the bigger cake. I know it sounds easy when you do not factoring in office space, number of employees, and tower/roof costs that only seem to go up. If you mind my asking what "slices" of pie do you offer?

I know some competition is easy to work with and problematic. I find it better to keep good relations even if it is competition it's better to reach out. If you have a lot of competition is it not a better idea for two, or three competitors focusing on securing areas of interest to expand or make it a every man for himself deal? I have more affiliates and only a few problematic competitors.

Overlap happens, I see other's bridges and ap's I try to avoid them. Do you use DFS channels? I use spectrum analyzation for each ap or repeater I generally find clean enough channels. Then there is always the relentless ones out there. I try to make peace the best I can and remind them that we all want the same thing. The approach I have tried is ask them who is the monopoly? Spectrum, FIOS, or another big company. I do not recommend this practice but if all else fails if you have the better tech it will take them more effort and resources to hurt your infrastructure.

It has been a long day and still working. 80 hours +/- a week. I need to rest these eyes. Someday I will get that vacation