Coporate WIFI Deployment - Cisco Hardware

[Donny_Dog]

I have been tasked with ensuring our 70 site network has WIFI added to extend operational usage.
Whilst I've spec'd and semi-played with the potential hardware, I am unsure on the correct approach for site surveying.

I've bought 3 X AP's that have been converted to autonomous (long term plan is use them in lightweight mode) so that I can quickly set them up without the controller.

I need some pointers on my methodology here:

• At a potential site, I've obtained estates plans and 'had a go' before getting there, where I think the AP's could go, based on walls, ceiling density and external noise factors.
• At site, I've set the AP's up in these 'guess' locations.
• I then use an Etherscope (Fluke Networks) to map the distance from the each AP, to where the SNR becomes a value less than 25 (at this point, I'd say the signal is OK, and we need to look at this being the boundary for this AP - is that right?).
• I then line up another AP so that the Signal stengths are acceptable and just slightly cross over using inSSIDer application (say -60db from one AP and -60 from another AP) and class that as being able to roam with a decent SNR calculated.
• I'll then crack open the 3rd access point and attempt the same process again, until I'm happy I've got a triangle of coverage where a user could roam and not lose connectivity (and still retain healthy throughput).
• I then mark this down on the estates diagram as potential locations for the AP's and move on.
• Is that the most logical way to go?

My equipment list to do a survey is as follows: (oh and they wont buy anymore kit..)

- Etherscope
- 3 x AP's
- Estates map (sometimes wont have this, so will draft my own )
- laptop with inSSIDer installed
- pen and paper

Any real world examples of how this is done? if you had the same equipment as me, would you deem the process as acceptable?

I've said we should not accept an SNR of less than 25. So far, early tests have shown noise in most buildings to be -90db which allows a minimum signal strength of -65db. Does that seem acceptable? They want to use b/g/n speeds through an array of different hardware.

 

[Cookie]

That's pretty much how I've done it at previous jobs (I didn't set it up here)

Massive site diagram, draw circles of varying colours out from the AP's, then move the AP's to ensure they overlap/you've got good coverage everywhere

What's the space you've got to cover?

The way we have it here is 1 AP in each office (not sure on the size of the office, but there are 100 desks or so) and then loads of repeaters dotted around. Perfect coverage everywhere.

 

[Donny_Dog]

That's pretty much how I've done it at previous jobs (I didn't set it up here)

Massive site diagram, draw circles of varying colours out from the AP's, then move the AP's to ensure they overlap/you've got good coverage everywhere

What's the space you've got to cover?

The way we have it here is 1 AP in each office (not sure on the size of the office, but there are 100 desks or so) and then loads of repeaters dotted around. Perfect coverage everywhere.

What repeaters do you have?

The space will vary based on each site. Some have 4 floors and are tricky to get RF signals, some are old school buildings that are made of paper. At least I'm doing something right then! Just need to research this repeater lark. I have only spec'd up AP's and the controller so far.

 

[bully]

Have you not got some software to plan the radio coverage?

When I worked with GSM cells, we had a prog which mapped it all out for us.

 

[Clart]

I've had a nightmare with RF coverage in our plant. Plus they need outside coverage in the Goods in yards - Even with Cisco 3502e APs and 3 aerials, the signal is just absolute dogshit.

 

[Cookie]

Try putting a wireless network in in a place covered in marine radar signals

It wouldn't surprise me if my swimmers didn't work because of that place

 

[Darren S]

Wireless - meh. I don't envy you that job one bit.

Is has it's uses, I'll give it that. But I resist and fight any decision here to implement it. Far too much hassle for very little gain.

When people have an issues about 'seeing wires' - they can simply GTFO as far as I'm concerned. Run a cable - click both ends in. It works.

D.

 

[Cr33do]

Sounds like you have thought of a methodical and entirely reasonable approach to me.

I did attend a site survey once, a couple of years ago, and we used a cisco AP mounted on heavy-duty tripod to ascertain positions etc. Cannot remember the software we have but essentially its a Cisco WIFI card that sits in a laptop and we had plans loaded into software. Place AP, walk around and the heat maps are drawn on to the plan No need to used 3 AP's at a time etc.

I cannot give you concise info I am afraid - I'm more experienced with wired LAN/WAN that wifi. Dont forget that wireless is a shared medium. If its the only method of connection you may need to up your AP count based on the head count in that area.

PS - I work for a Cisco Silver Partner; if you need any kit sourcing or require outsourcing your wireless survey PM me

 

[Clart]

Wireless - meh. I don't envy you that job one bit.

Is has it's uses, I'll give it that. But I resist and fight any decision here to implement it. Far too much hassle for very little gain.

When people have an issues about 'seeing wires' - they can simply GTFO as far as I'm concerned. Run a cable - click both ends in. It works.

D.

Mobile scanners on cables wouldn't go down too well in most places.

 

[Donny_Dog]

Just done one wing of a building as a demo using the methodology i explained.... and.... WIFI is definately not FTW.
Its going to take ages and is supremely boring. At least my idea stands to reason, but I think I'd rather it didnt.
Death by access point.

 

[Cr33do]

Good luck mate - you do have a big job ahead of yourself!

 

[Donny_Dog]

Good luck mate - you do have a big job ahead of yourself!

Cheers