Wireless Coverage

Unanswered Question
May 16th, 2008


I need some information about Wireless deployement in a huge surface.I want to cover a surface of 40 ha that contains about 400 appartements and Villas.

Is there a solution to do that with a few number of Access Points?


I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
Alex Pfeil Fri, 05/16/2008 - 03:45



You need to determine how good of coverage you are looking for. How many users you are going to have. The best way of determining how many access points you are going to need is to do a site survey. If all of the appartments and villas are the same. You could install a few access points and test the coverage. Move the access points as far away as possible as you test. If this is not feasible, you can have a company come out and do a wireless survey for you. There are different software products out there for site surveys and wireless planning as well. But, It might be cheaper to have somebody plan it for you. Just remember, if you use good access points (like cisco), they are going to be better at handling traffic and have a greater range right out of the box. Different access points use different antennas and the placement is vital. Hope this information is what you were looking for.



ericgarnel Fri, 05/16/2008 - 07:35

Alex is correct, a survey is essential.

Also, define your goals for the wifi network

Are you going to provide wireless to the outside areas or do you want the signal to penetrate the interior? You will have to survey from each level of the villas as the signal will not penetrate equally as the height increases

Do you have set policy for bandwidth and acceptable usage?

Have you set the expectations for what the wireless environment is like? It is much easier to appease end users if they have an idea of what to expect in terms of coverage, access times and speed

ismail884 Fri, 05/16/2008 - 09:11

thanks for your reply.

I've do a site survey with Cisco AIRONET 1100. for the quality of Signal I look for "Good" or "Exellent" strenght.

the result of the site servey was:-implement an AP in each Appartement and 3 APs in each Villa, so it's so difficult to implement a Wireless Solution with this enormous number of APs.

have you any idea about another solution that permit to cover this surface.

many thanks


ericgarnel Fri, 05/16/2008 - 09:20

The use of directional antennae would help.

Also, are you going to provide a as well as b/g?

Instead of using "good" or "excellent" as a measure, use the actual radio strength and throughput tools like iperf to measure quality.

You will also want to measure for interference and channel placement and overlap.

What radios/antennae are you going to use in the deployment?

ismail884 Fri, 05/16/2008 - 09:32


the goal of the solution that I look for is to provide Internet Access to all resident in this surface via Wireless.

for the type of Radios/antennas, I look for something that cover the almost possible surface.

with indoor AP I calculate that I will need about 500 APs.

have you any Idea about the Mesh solution with the AIRONET 1500 Series?


ericgarnel Fri, 05/16/2008 - 09:44

The Mesh solution with the 1500 series is very nice, but also very expensive. You may come out even in $$ with the indoor radio deployment.

The # of radios+antennae required will depend upon what service level and area you want to cover.

How did you arrive at the 500 AP #? Are you providing signal down to the lowest bitrate or or are you basing your measurements upon highest bitrate?

ismail884 Fri, 05/16/2008 - 10:06

when I did the site survey, I was in the obligation of installing the AP inside the appartement to have a good Signal,so 1 AP in each appartement( 1 AP * 380 App) and since the Villas have a huge surface I noted that I need to implement 2 AP in the RDC and 1 in the 1st Floor, so each Villa will have 3 AP (3 AP * 45).

ericgarnel Fri, 05/16/2008 - 10:17

Is the obligation to have an AP in each Apt? or just good signal. Depending upon construction materials and use of directional antennae, you can provide coverage. Here is an example

Floor3: |[AP>][>>>][>>] |

Floor2: |[ <<][<<<][<

Floor1: |[AP>>][>>>][>>] |

> = directional signal

[ ] = floor

Alex Pfeil Fri, 05/16/2008 - 10:22

I would like to add that you can use 25 foot extension cables for those antennas. We have used those for better coverage in areas of bad signal propagation for better coverage.



ismail884 Fri, 05/16/2008 - 10:29

the obligation that we have is to obtain a good signal in each appartement, the construction materiels make obstacles to signal.

so when I put the AP outside to cover more than 1 appartement I note a very bad signal inside.

Alex Pfeil Fri, 05/16/2008 - 10:34

I would try installing an access point into a hallway, and running an antenna into two different apartments. Use a 25 foot extension cable and run one cable into one apartment and one into another. Therefore you would essentially get the coverage and be using one access point to cover 2 apartments. You could also try installing an access point in the attic in between 2 apartments as well.

SHANNON WYATT Fri, 05/16/2008 - 17:08


An access point has two antenna mounts for a specific purpose. The purpose is to combat the effects of multipath. When using two antennas you need to keep the antennas in the same area. Seperating them out will look great on a survey, but will break on load.

Solutions like leaky coax can be used when you need to cover a large area but few users, but remember, these things are hubs, so you have to keep in mind the limitations you would put on a 10MB hub.

When you are talking WiFi, you are talking about radio signals. The only way you can increase range is by increasing signal gain or transmitter power. But as you increase gain and power, you also increase the area of coverage that will directly affect neighboring access points. (the exception is a leaky coax system)

When designing a wireless network you should be designing it based on the application and clients, not how to get the fewest number of access points. That is a sure way to lead to a poor design.

ismail884 Sat, 05/17/2008 - 02:38


To provide a solution to a client, it's so important to take in consideration the solution's cost. After the site survey that I did,and with results that I obtained, I could not provide a well design with considering Cost and easy management of the Wireless Network. It's not 30 or 40 APs, I talk about 500 APs to install to cover all the surface (40 ha).

SHANNON WYATT Sat, 05/17/2008 - 08:13

True, but if the solution requires that kind of density, than that is what it requires. Deploying a bad solution is usually worse then no solution.

You probably should think outside of the box and look to deploy something else (Ethernet, LRE, DSL, etc), and deploy limited WiFi to the people that want/need it, and charge back appropriately.

ismail884 Mon, 05/19/2008 - 03:38


I thinks that the best solution that I can propose is to deploy Ethernet cabling inside each building and deploy Wifi outside to cover gardens and outside surface.

thanks for all


Alex Pfeil Mon, 05/19/2008 - 03:31

If you only need to cover two apartments, 2 to 4 computers sharing a Wireless access point will work fine. I don't understand why you would disagree with this especially when the connection to the internet is probably not more than 10Mb/s.

ismail884 Mon, 05/19/2008 - 03:53

The project is for a hotel that has a surface of Ha of Appartements and Villas for rents, So they need to provide Internet Access to clients every where they move in their Appartements, Villas,Gardens,Piscine.for outside we can cover an important surface since there are few obstacles, but inside buildings it's so difficult to cover more that appartement with an access point.

Alex Pfeil Mon, 05/19/2008 - 05:09

I think that if you place your antennas correctly, you will be able to cover at least two apartments with one access point with no problems. This fixes the issue with having to use one access point per apartment.

nrparks024 Sat, 05/17/2008 - 16:15

another possible solution would be xirrus arrays. google xirrus.

SHANNON WYATT Sat, 05/17/2008 - 17:33

Xirrus has a nice story, but the reality falls short of the marketing. You could put enough gain on an antenna to boil water, but once you hit the inpenatrable object the game is over. Their products probably make more sense in places that you have to cover a lot of users in a small area, like a large cube farm, and you want to reduce the number of network drops.

nrparks024 Sun, 05/18/2008 - 08:36

It is still worth looking into. The Xirrus guys I dealt with provided a free survey based off their device of course, and let us test 2 arrays . That was a year ago. We haven't had a complaint since. Building is 15 to 20,000 sqft made of brick and concrete and mostly offices.

dennischolmes Sun, 05/18/2008 - 11:40

I am interested to learn morefrom an actual Xirrus end user than their marketing info. From an RF science standpoint I don't think their arrays can support any more users than Vivato and Vivato was limited to 50 concurrent users. Is this what you are seeing or is your number of users less so you just never notice the limitation? I would be interested in learning just how they eliminate the co-channel and adjacent channel interference as well as scalability issues. Again, no marketing info please. I would like first hand user data from you. What data rates connected at, ranges, attenuation info, etc.

SHANNON WYATT Sun, 05/18/2008 - 13:18

From the Xirrus website it seems that they are stressing 5 Ghz, which makes sense for a lot of reasons. I've been telling my customers the same things for years. But the reality is that in most deployments you do not have a lot of control on the client end. In hospitals you have to support 2.4 GHz wifi badges, in manufacturing you end up supporting 2.4 Ghz WiFi scanners, and in education you have to support every piece of WiFi equipment since the inception of WiFi. It is great that the Cisco 7921 supports 5 Ghz as voice was always a pain in the past, and now is much easier.

It looks like the Xirrus uses multiple radios with sectorized antennas. You could have a setup with very good results in the 5 Ghz band with that configuration. The problem in the 2.4 band is that even if the antennas on the access points had zero channel overlap, the clients have omni directional antennas. So the clients would be the cause of the same channel interference. Rememmber, each client radio is essentially the same as the access point radio, and you have little to no control over that device in most deployments. Sectorizing the antennas make a lot of sense in a very large open areas, but most buildings tend not to be that large or open. In this case that is certainly true. Now if they only have three 2.4 Ghz radios the problem with the same channel interference goes down.

CCX is one way to bring a level of control to the WiFi network clients, but that requires CCX client software. When using Cisco, Atheros (same as Cisco), Intel, etc, CCX capable device you have to use the manufacturers' supplicant to get CCX on the clients. The problem in an a lot of deployments is the CCX client limits the client side experience (Group Policy, Login Scripts, etc.) So without CCX the clients are out there blasting away with something between 2 and 4 DBI gain on something between 30 and 100 mw. Again, these are omnis (still waiting to see a laptop with a rotating parabolic dish!), so you are pushing out decent amount of signal a good distance with these clients, potentially into the area of another access point or clients on the same channel.

This is why when designing a WiFi network you need to look at the application(s) first, the clients second, then look at the access point and antennas as a way to provide the necessary performance for the applications to behave and the clients to have a good experience. Hopefull the advent of the 802.11n will make this easier as we will have many more channels to work with. But until we get to the point that a decent number of clients are .N we have to deal with the issues of the 2.4 Ghz spectrum.

It seems like this deployment is for apartments. It may be possible to deploy wireless outside the units in open areas (maybe the mesh access points) and get coverage in the units. RTS/CTS would need to be enabled, but if you only had to support web access that may work. The issue would be how far inside the buildings you would penetrate. But from the sounds of the limited survey that was performed it may not give the desired results.

If the goal is to provide internet access to the units they may want to provide internet with an ethernet handoff (ethernet, LRE, DSL), and let the renters/purchasers of the units deploy their own WiFi if they choose to. That or offer it to those that would like managed WiFi the ability to purchase it for a fee.

dennischolmes Sun, 05/18/2008 - 14:04

You make all of my points exactly. I am wanting to hear the results from an actual Xirrus customer. I am a strong believer that Vivato, Xirrus, and others are simply playing on the lack of knowledge that abounds about wifi, its limitations, and its good points. As I sit on the IEEE standards association and have had constant input on the standard, I fully believe that 802.11n in 5ghz will be the way to go. I just wish these start-up companies would quit preying on folks who have not gotten a lot of education on how to shop for a enterprise grade wifi solution. As for the level of control you speak of when dealing with clients may I point you to 802.11k and 802.11r. K has more to do with client information and R is standardizing roaming feature sets.

SHANNON WYATT Sun, 05/18/2008 - 16:03

No doubt. Client control has always been a weakness in WiFi. It was a bunch easier for us a few years ago when we would sell the customer the access points and the PC Cards to use with the network. I'm glad that finally we are getting to the point that I can start dropping support for plain old 11B!

ericgarnel Mon, 05/19/2008 - 07:32

Having used Vivato panels in the past, and seeing the results of a "shotgun" approach to coverage, I totally agree with Dennis.

I have also seen the xirrus solution up close and had a chance to inspect the innards of one and was not impressed.

Here is something else to consider with "big box" solutions: if you lose the device and/or power to it, you lose a lot of coverage, whereas a high density cell design will compensate for the loss of an AP or two.

Y'all are striping the primary & secondary controllers to adjacent APs I hope!


This Discussion