Decrease font size
Increase font size
Topic Title: Large scale WiFi on campus and commercial area
Topic Summary:
Created On: 26 October 2011 02:26 PM
Status: Read Only
Linear : Threading : Single : Branch
Search Topic Search Topic
Topic Tools Topic Tools
View similar topics View similar topics
View topic in raw text format. Print this topic.
 26 October 2011 02:26 PM
User is offline View Users Profile Print this message


Posts: 2
Joined: 31 May 2011

What's the usual tech specifications used in building a campus wide or commercial WiFi?(eg. 2.4 GHz) And whats the difference between the devices set up as access points and our home use wireless routers? In fact, I'm still not very clear what to ask about, just want to see how the large scale system differs from home use wireless system. Where shall I start to deploy a 2,3 floors coverage wireless system if I have only the knowledge of setting up a home use wireless router.
 26 October 2011 03:29 PM
User is offline View Users Profile Print this message


Posts: 66
Joined: 04 May 2009

The two big issues you have with large scale deployments over home systems are mainly ones of signal interference and client density. Building construction can sometimes play a part as well although it's usually relatively easy to design around that.

The signal interference issues are because you only have limited numbers of non-overlapping channels in both the 2.4 and 5GHz bands and you need to ensure that APs are laid out in such a fashion that the signal from neighbouring devices do not interfere with each other. This is made worse in 802.11N where channel bonding is used.

Client density comes from the fact that even the APs from the top brands in the market cannot deal with too many clients associated at any given time before they stop being able to process traffic effectively so if you have areas of high density you need to consider adding more APs to handle the number of clients in that area but then you fall foul of the first problem!

In the initial stages of planning it's critical to conduct and RF survey within the parameters of the desired signal coverage - this is to make sure you know where to place your access points to get the appropriate coverage with no dead spots. This also helps you sort out the logistical issues of getting data cabling and power [or PoE] provision to the mounting point for the APs.

When it comes to the actual hardware most of the large scale stuff these days is controller based. The access points themselves have limited or zero touch configurations, unlike a home device, and instead they pair with a centralised controller which acts as the configuration engine. The advantages of this are obviously that you can perform configuration in a single place and push it out to hundreds of access points without needing to visit them all individually. The controller also takes on the channel management role and will come up with the best arrangements of channels to avoid interference without you having to manually plan then - it's not perfect but it does the job most of the time. The controller can also provide some load balancing by making clients associated with an AP re-associate with another AP in the area to balance the load across a cluster.

Finally in these controller models the traffic is usually tunnelled from the APs - which support multiple separate SSIDs - back to the controller where they are mapped onto VLANS on the backhaul network. You don't have to do this but the advantages are that it provides a nice central choke point for firewallling, guest access control, content filtering and other things you might want to do to provide security.

The actual configuration on a per-AP basis is not that different from a home unit, mainly the difference is in how they all interact with each other to provide a coherent infrastructure. There are solutions from some manufacturers that then provide unique features to add something to speficic situations. Cisco for example pushes their clean air technology - which is basically automatic channel reconfiguration to avoid interference sources and Meru have a token-ring like technology that allows their APs to deal with extremely high numbers of clients associated and also allow all the APs in a cluster to operate on a single channel.

In terms of where to start the first thing I would do is start from drawings of the building and identify viable locations to place access points. From there place a cheap AP at each position in turn and using something like wi-spy, wifofum or wirelessmon map out the extent of the signal coverage that each location gives you to ensure you have no dead spots. Then identify any areas that might need higher client density than other areas to know where you may need to add extra APs. You'll come out of this knowing a good idea of how many APs you'll need to budget for - I would normally add a few extra in when asking for budget just in case.

Once you've got that you can look at manufacturers and get costings and then it's a matter of purchasing it, rolling it all out and plugging it in.

The above largely applies to wireless networks for data usage, if you want to offer wireless VoIP or location based services the rules change somewhat.

Hope this helps

Jake Greenland, CEng MIET.
CCIE #22595
 26 October 2011 09:48 PM
User is offline View Users Profile Print this message


Posts: 2
Joined: 31 May 2011

Thank you so much. It's exactly what I tried to learn about!!!

See Also:

FuseTalk Standard Edition v3.2 - © 1999-2016 FuseTalk Inc. All rights reserved.