Is This Feasible? Congested, Non Line-Of-Sight City Wireless Network?

We were asked to find a wireless solution for IP Survellience in a narrow congested, serpentile area spread across 5 Kms. They cannot dig and lay fiber or Coaxial Cables(for EoC). In such environment, with difficult Line of Sight for Wireless Systems, is there a Solution?

I'm tempted to write "unfeasible". So I thought about consulting the experts here. Is there a Solution?


Hi 1!

I think it's feasible because regardless of the obstructions, as long as you can see the sky you could always use a VSAT setup:

Cisco's VSAT service for instance, provides up to 10 Mbps per dish:

Completely IP-based, the Cisco IP VSAT Module supports the latest in satellite technology with respect to access scheme, modulation, and coding to provide maximum efficiency and performance from the satellite network. The module is compatible with satellite services that use Gilat SkyEdge hub systems. The module is the indoor unit (IDU) built into a network module form factor, and it connects to the outdoor unit (ODU), which consists of the dish antenna and the transmitter/receiver using coaxial RF cables

Now it might not be the best way, or the cheapest, but it's one way.

To help figure out some other ways can you describe the topology of the site some more, and provide basic requirements and budget info?

Thanks!

you gave me a great idea with the VSAT Solution.

The topology is isolated PTZ cameras on Lamp Post streaming to a point. So Multi-Point to Point(Star). Think of it like you are in your office and you want to capture wirelessly from a CCTV which is obstructed by houses, hilly city developments. Vegetations.

Budget is US$25,000 for Networks. (the tragedy of everything).

VSAT is definately not a solution for PTZ cameras (latency). 5GHz point to point is generally the way to go, which may result in a few repeater sites to get line of site to the cameras. This all depends on whether you have access to power poles or other locations along the route. With Ubiquiti NSM5-US ($75 each) you can accomplish this. If you only have access to the headend site and the camera sites, you would need a longer wavelength, like 2.4 GHz, to get through the obstructions, using a multipoint or PTP scenario. It is not stated where this is located, so in urban areas, interference can be an issue using 2.4 GHz (WIFI). For 25K, I also did not see you state how many cameras were to be deployed.

Using 5GHz PTP, I have been quoting probably an average of 4-5K per HD ptz camera (cities install the power harness that I give them to the utility poles). I would expect a repeater pole (two radios, no camera) would be about 1-2K (mostly labor and lift charges on this). Ubiquiti is neat because it has a second NIC port with POE out, eliminating the need for a switch. I have gone out to 5 links using this method with no loss of frames or video from 1.3 MP 15fps PTZ cameras. You need to be aware to program dedicated frequencies for these links using declining bands (for example, 40 MHz first link, followed by 20 MHz links in 5.8 or 5.3 GHz links, with maybe 10 MHz bands for final legs). Look at the spectrum and provide at least 5 MHz separation between bands (for example, using 5765/20, this would occupy 5755-5775, so the next band may be 5785-5805). Do a frequency map/plan on this. You can reuse frequencies as long as they are not line of site to another link that may cause interference. Do not just use "auto" mode on the radios.

Pull up a Google earth shot of the area and camera locations, and see if you can "line of site" your way to the cameras. Include repeater sites where needed, (as long as you can get access to the site).

You could also try 900 MHz, which will work non-line of site, but bandwidth is limited as is quality of service, which may/will interupt a H.264 video stream. (You could do a combination of 900, 2.4, and 5 though with proper planning).

VSAT is definately not a solution for PTZ cameras (latency)...

Of course we only found that it was a PTZ after I asked for additional details, so I have no problem saying that latency could be a big issue with VSAT. Anyway, the OP asked if there was ANY solution, note that I also said it was probably not the best solution.

Rather it was to say there was at least a solution, without line of site, without 3G phone service, without 'multiple repeater sites', without a spectrum site survey. As for the PTZ latency, with UDP and depending on the Satellite, you are looking at around 700ms one way, min. Call it a second and a half from the time you click til the video moves. Terrible for trying to track a fleeing suspect. But not so bad for repositioning to a new ROI, or using presets and tours. Or camera-side object auto-tracking.

Curious, what kind of latency have you seen with your 5 hop setup?

In any event, what I hear you saying is that if you have access to poles etc, then multi-hop is the way to go, and then I like it as well. And I think Ubiquiti makes cheap, powerful and reliable equipment.

I considered 4G as well but I didn't want to assume there was high-speed cellular without knowing. Let's see what other details the OP can provide.

What about using cellular? Latency is an issue but cellular networks are designed to work without line of sight, have broad coverage already deployed and increasingly are capable of handling high throughput levels needed for video.

Btw, also perhaps Whitespace Wireless for Surveillance

With a 25K budget, that's completely doable, and working within that budget you can shift installation & mounting to the city resources... You need a good wireless distributor with the resources to help recommend the right radios. Radios that you can learn and become comfortable and self sufficient with. Lots of good information here http://www.wlanantennas.com/faq.php

For best reliability, (we get 100% connectivity in the worst of lighting storms) and to get that you need a secure frequency hopping protocol. Proxim calls their WORP, and whatever radios you get you need their equivalent. (non of these protocols are compatible)

...and working within that budget you can shift installation & mounting to the city resources...

Andrew, just curious, what does that mean exactly? Make the city pay?

In many projects with tight budgets we shift the burden of installation and aiming of cameras and radios to provided employees.

Then what do you do? That's most of the work of a project. Just commissioning and configuration then?

Call me crazy - but we've never replied to an RFP. We only do projects where the clients wants us to provide the solution and our mandatory RMM/SLA support agreement. We know the budget and the criteria including the needs requirement, recommendations, design, installation, integration, training, maintenance, and support. Therefore we have no issues supervising installations by client provided employees to keep the total project under budget.

It would be helpful to know more, how many cameras address the area? what is the expected bandwidth. Can they install, or do utility poles exist in the area?

Depending on the requirements, the answer might be to use backhaul radios using a FHSS (frequency hopping, spread spectrum) secure protocol (Proxim Worp, MiKroTik Nv2, ubiquity AirMax) for the link and at the far end, us 900mhz NLOS APs) 900Mhz is a busy spectrum, and if the area is heavily populated that may not be so good. Talk to a good wireless distributor about a neighborhood cell solution.

Thank you Andrew , I found a good case in wireless network of a Campus. The Product Standard of the one I saw of Wireless Manufacturer is very effective but does not give high availability without a Fiber Backup. I was hoping to find people who have sufferred through rain, hail, lightning and sun to implement a robust solution. Unfortunately, wireless has its limits.

Does not give high availability without fiber backup.

How high does it need to be?

Is a mesh network out of the question because of budget?

My mesh network supplier tells me it'll shoot the budget. Any lowcost manufacturer on mesh, otherwise 802.11ac is also good, no?