Proximity is King: Dispelling two myths of public access WiFi

I often respond to questions related to the best way to deliver a public access WiFi service for various applications including hotel rooms, shopping malls, and wider outdoor areas like parks gardens or tourism districts.

 

There are two big misconceptions about delivery of these kinds of services that invariably result in a poor outcome for all parties involved in a project implementation.

 

 

It is tempting to think of WiFi like a sound system:  “Mount the Access point high above the crowd and use the most powerful transmitter you can find”

 

There are two good reasons why this is NOT always good idea, both related to the important second party to every WiFi service: the client device.

 

Apart from the fact that transmitter output power is limited to a maximum by government regulations, higher transmit power on the base station does not always lead to a better service to the customer.

 

The first reason for higher transmitter output power not always being such a great idea is that the base station is only half of the system – wireless networks are a two-way operation: the client needs to ‘hear’ the base station, but so too does the base station need to ‘hear’ the client.  Sure, you can flaunt the regulations and deploy an Access Point that transmits at 5 times the legal limit, but the client devices will all be transmitting at (or, usually, well below) the regulatory limits.  The end result is that the customer can see your network SSID when scanning for WiFi, and it is detected as high signal strength, but when they attempt to connect, they get either a poor connection, or the device fails to connect at all.  The outcome is that the customer assumes there is something wrong with the service, or experiences unsatisfactory performance and your service reputation suffers significant damage.

 

Secondly, many modern WiFi capable devices, smartphones especially, adjust their wireless output power to save on battery.  When the device receives a strong signal, it will often reduce its own transmit output on the assumption that a stronger signal from the base station mans that it is closer.  The result, of course, is to exacerbate the ‘reply signal’ problem and further reduce the effective network connection performance.

 

 

To address the limitation of client transmit power, a common idea is to simply use a super high gain antenna.  An antenna acts a bit like a ‘signal amplifier’ for a wireless device – both on transmit and receive.  A higher gain antenna will drive the transmit signal further, and will allow the base station receiver to detect more distant, weaker transmitters.

 

This may appear to be a sensible solution to the problem of range, but there are three very good reasons why this will rarely produce a good result:

 

 

It is easy to forget that WiFi is a shared medium:  All wireless networks share a common radio spectrum, and there is limited space to go around.  In fact for typical WiFi services, although there are 11 available channels* there are only clear channels available: channels 1 (2412MHz), 6 (2437MHz), and 11 (2462MHz).  That means that when there are more than three wireless networks operating in the area, there will probably be other networks, other base stations and other client devices competing for access to the same spectrum that you are using for your own services.

 

Although WiFi protocols are designed to cooperatively co-exist with other wireless networks in range, transmissions that are not EXACTLY the same channel cannot be recognised as a competing wireless network, and so essentially becomes nothing but noise, or interference, to your own network.

 

With this in mind, consider the effect of a high gain antenna in a public access environment.  Sure, the base station will be able to detect wireless client devices further away, but it will also detect other base stations, and clients of other networks farther afield!  Therefore, the higher gain antenna that you use, you end up sharing the available bandwidth with a greater number of other networks, or (worse) picking up more interference from other transmitters that are either not using one of the clear channels, or are not WiFi transmitters at all.

 

 

Another important reason that a high gain antenna might not deliver the expected result is an inherent limitation of the WiFi protocols known as “Hidden Node” problem.  For a detailed description of Hidden Node, there is no better article that I am aware of than this one: http://en.wikipedia.org/wiki/Hidden_node_problem

 

A higher gain antenna may extend the reach of your base station, but it also increases the maximum potential separation of the participating clients.  This increases the potential for hidden node problems to affect your network operation.  Keeping in mind that increasing numbers of hidden nodes on any wireless network causes an exponentially devastating effect on the network performance! 

 

 

A third consideration that is often overlooked when selecting a high gain antenna is the shape of the ‘beam’ formed by the antenna.  Due to the way antenna design works, as antenna gain increases, it becomes more ‘directional’.  An antenna with gain of ‘1’ dBi, for example, forms an approximately spherical transmission**.  So the signal strength will be the same when detected by clients the same distance away from the antenna in any direction.  If you could add a reflector to that 1dB antenna to turn the spherical shape into hemispherical (‘half a sphere’) then all the transmission energy of the reflected signal will effectively double the ‘power’ in the remaining half-sphere, or gain of approximately ‘3’ dBi.

 

So a lower gain antenna will produce a beam pattern that is closer to spherical (in reality a bit more like a fat donut) so that when the base station is 3-4m on a post, the signal is equally well received directly below the antenna as a few meters away from the base of the mounting post.  As the antenna gain increases, the reception directly below the base of the pole decreases, and the signal is concentrated more toward the zones farther away from the pole.  As the antenna gain increases, the beam shape of the antenna is less like a fat donut, and more like a flat pizza.  When the gain becomes very high (12dBi and above) the beam shape can be so flat that the wireless transmission shape (and reception zone) becomes so flat that the signal essentially passes straight over the heads of the intended clients, thus resulting in very low signal throughout the intended coverage area.

 

Our own experience, as well as feedback from customers and partners, suggests that an optimum antenna design for outdoor areas is an 8dBi antenna mounted approximately 3-4m high.

 

Although antenna gain and base station transmit power are rarely an effective way to increase the range of public access networks, it is worth noting that there are some special cases where these effects will produce the expected result:

 

Transmit Power: when the client device is also controlled by the network operator (i.e. fixed wireless – WISP), both client AND base station transmit power can be increased to produce greater range.

 

Antenna Gain: when there are no other wireless devices operating in range, and only one client is expected to connect, higher antenna gain can help to improve overall range of the wireless network.

 

Neither of these cases, of course, are relevant in the case of a public access WiFi service, and so not particularly useful in thus context.

 

So in order to deliver an effective and efficient wireless network for use in a public access environment, we often refer to one primary concept: Proximity is King.

 

The best way to deliver a high performance data service across a wide area is to make sure that the client is always as close as possible to the base station.  A distributed approach to building a wireless network brings several advantages:

 

• You can reduce the transmit power of all access points, thus require clients to be closer, and thus minimising chance of hidden nodes

• You can use a lower gain antenna, thus deliver a more spherical beam shape

• The maximum available bandwidth available to the wifi channel is shared among fewer clients = faster speeds for all

 

There are a few different ways to design a distributed wireless network, some using cables to connect all the access points, and some using wireless.  Although a discussion of these methods are beyond the scope of this brief article, if you have an interest in exploring these concepts further, feel free to contact us to discuss in more detail!

 

* although Australian regulations permit 13 channels for 2.4GHz WiFi, there are two main reasons why we always recommend to adopt the international scheme:

 

i) If your service is likely to be accessed by international visitors, their devices will be configured to only detect and access 11 channels – thus if your network operates on channel 12 or 13, it will be inaccessible to international devices.

ii) Because of point (i), invariably when there are existing wireless networks operating, they will be almost certainly already configured to match an international 3-clear-channel scheme.

 

** note that this is a highly simplified description of antenna design, for conceptual discussion only!