08-10-2017 09:53 AM
We currently provide rural service to a city of about 4k residents. We use a grain elevator located in the middle of town that has great LOS to the outskirts, where we have several repeater sites, but trees prevent us from being able to provide ubiquitous coverage in town. Noise floor is around 100db and as far as I know there is only 1 other WISP in town and they only cover a small corner that is able to see their tower 20miles away. There is also very spotty cell servic. Recently the city approached us about becoming the primary provider after getting quote shock from several other companies.
In order for us to use UBNT gear we would need to spend about 100k just in gear to build out all the access points needed to ping pong coverage around all the trees. I imagine that we would use up most of the airspace just with the shear number of radios needed.
The more I research the more it seems like 4g would be a viable option for around the same price. The town is roughly 4miles x 2 miles with our tower being located pretty much right in the middle. It is a multi seed sorting facility and has 3 towers that are roughly 80ft off the ground and have good spectrum insulation in the form of large metal buildings that separate them from each other. Using all three of them would give us a great 360 of the town with space for 60-90 deg of overlap without interference worries
Assuming that each house has 4 members I am looking at around 1k houses of which I would need to cover about 750 with a 4g, and the rest with UBNT or Combinations of both. Looking to provide around 20mbps on the high end and 5 on the low
What are your experiences with current 4g tech? Does it work as advertised, or is it more of a hassle then it is worth. Do you think a solution like this is viable given the above details? Do you have a different company or solution that you might reccomend?
08-10-2017 10:48 AM
You are working with only about 50mhz of spectrum today. That is barely enough to do 2 APs. To add insult to injury, when you factor in multipath and NLOS issues, that 50mhz has poor spectral efficiency so you won't get as much throughput as you would like.
Sorry to bring bad news.
08-10-2017 01:14 PM
Are there any viable NLOS solutions? Right now we are running a 900 Mhz sector that is performing pretty well given that its a 900. However we have residents looking for higher data limit packages and I'm kinda stuck as to where to go next.
08-11-2017 06:22 AM
There is only 3 useable channels but I would suggest 2.4. We have some success in the city at 10 wide, otherwise you can go 20 wide, just hide the ssid I use a rocket 2 AC, and have had good luck with them so far.
08-11-2017 06:55 AM
I would look at the microcell pop design using 5GHz AC for AP's and a few M2 @10MHz for worst case use only. In my wisp I use a mix of M5 / 5AC / M2 / LTE. As others mentioned LTE is no magic fix I do see much better performance vs 2GHz. Part of that is hardware as you are comparing a $75 rocket to a $3-8K base station and the lower noise floor on the 3.65 band in my area.
That being said in my experience the 5GHz AC is the preferred install option with the best spectrum effiency in current PTMP products followed by LTE then 2GHz / 900MHz.
In your case I see the mini-pop as the only real solution due to density / trees / and bandwidth you are looking to provide. Maybe a mixed design consisting of micropop and might I add GPON fiber would be in order. As the city wants you are the provider they may be able to help with easement / pole access?
08-11-2017 08:15 AM
I guess I should add some details.
We currently have 2.4 and 5 ghz sectors up on the tower, along with a 900 that we call the "junk sector"
For the people who can see 5 ghz works just fine. We are also using 5ghz to backhaul several micro pops.
We use 2.4 as a go to for NLOS and for some micro pop backhauls with small amounts of trees.
900 is used for last ditch effort for people who need less than 10mbps and are desperate.
The problem that I am currently having is this -
With so many houses stacked so closly together and almost all being paralell to our tower instead of perpendicular. Once you get 4 or 5 houses on the same side of the street 2.4 becomes unreliable for higher speed clients. I havent tried 2.4AC or AF yet, have a pair of both in the mail, but generally speaking there is not enough BW to handle more than 3-4 homes with a 2.4 BKHL / Micro Pop situation.
Same issue with the 900 sectors.
Cities around here are broke. We already have a hard enough time just doing service. We only use private property and therefore don't generally pay franchise fees to the city. Part of the reason that there are no providers in town is because the city basically put up a barrier to entry with their taxes, and that is the reason that companies are quoting so much for fiber install. 2 Members of the city council, who are also our customers approached us about providing a solution because they know how we operate and thought that we could at the very least fill the gap until the city comes to it's senses.
So let me reframe the question.
We are already running a mix of UBNT gear on our tower, but are needing something to fill the gaps where we are struggling with the UBNT products because of noise, or NLOS issues. We currently cover about 100 of the 1000 estimated houses.
Would it be feasable to use LTE as my backhaul links for micro-pops to areas that have poor LOS? Possibly one CPE at each end of the street and then UBNT AC Lite Sector and NB19AC to connect the houses to the LTE BKHL. In that case I would be using a mix of 5, 2.4, 900, and LTE as my backhauls, and then 5 or 2.4 AC for as many CPE's as possible.
I also heard that Cambimum 900 and 2.4 gear performs better under these kinds of conditions. Looking at the 900 base station and antenna cost, I am wondering if it's possible that the performance boost will justify the cost.
I understand that I am probably approaching the point where we will have to start using a mix of gear and manufactures to achieve the desired results. I just don't want to invest a bunch of money in a mediocre solution. QOS is the only thing that makes us any different from CC or CL and I don't want this project to blow up in my face should I choose to follow through on it
08-12-2017 06:21 AM - edited 08-12-2017 07:14 AM
Sounds like the city needs to get real here. I'm sure they saw a cash cow coming down the street when they set up the franchise fees etc, but the reality is that nobody is going to do this and pay them a lot of money any more.
On your end I think you need to approach this from the business end more than from the technical end. Make it clear to the city that you are willing to work with them to deliver service, but you don't have deep pockets and you're not going to risk your business to make it work without help from them. Get them to give you free access to poles or easements so you could lay fiber to selected areas where you can then deploy micro-pops covering out maybe 1-2 km. Do this in selected areas where you know there are customers. rinse and repeat. Or use wireless backhaul from the centralpoint to the micro-sites, and then plan to convert the backhaul to fiber in the future. That preserves the spectrun long term. Use GPS on AF or AC and you can reuse channels for backhaul. Only use GPS AC for the micro-pops and use LBE-AC for CPE to keep costs low. This can be done. Including with all UBNT gear - we do this in a major Metro area and in rural areas 60 miles away right now.
True story - several years ago we set up a system in a town a little bigger than you are talking about, using their water tank for a site. When the town right next door found out, they asked if we could do the same for them, including connecting their buildings etc. and we said sure. About a year later they came back and said they were ready to do a deal, and would only charge us $5K per month to use their tank. Once we stopped laughing (not really, we were all raise beter than that) we explained that we would trade the site for their connectivity, but we couldn't pay them anything. They went away, and we did too. Forward to 2 months ago and we get a call from them - Comcast finally said they couldn't/wouldn/t do anything for them, and they would now like to talk about us coming in and doing the trade. We said sure, but we're really too busy right now, but we'd be happy to see what we can do in September or so... Working on the lease agreement right now.
So do what makes sense for you economically, and show them it's in their interest to work with you, and that you're not Verizon or CL who will throw money at them (although as they've found out that ship has sailed). It's surprising sometimes what you can work out if you're up fromt with them...
"Humans are allergic to change..They love to say, ‘We’ve always done it this way.’ I try to fight that. "Admiral Grace Hopper, USN, Computer Scientist
08-12-2017 11:37 PM
I'll admit that I skimmed this pretty quickly so I definitely may have missed something, but I didn't see anyone mention the obvious AirFibre using 24GHz, or some of the competition's 60GHz stuff, for your backhaul needs.
Just thought I'd throw that out there, even though it seems like I have to have missed something.
Sorry if repeating!