Wireless Access

I'm trying to overcome some design challenges in a high density multi-floor office.  VoIP is in use so I'm shooting for 2 APs at -65dBm, 25 SNR, with a max of 2 APs overlapping at -85dBm, all on 5GHz.  After designing it in ESS, I could tell that I was going to have some issues with CCI.  I've since deployed the APs (225s) and did an active survey, and verified I've got high CCI on most floors as a result of APs bleeding above/below.  I've done everything I know to do to minimize CCI:

• use walls to attenuate the signal
• lower the transmit power while still maintaining the RSSI/SNR requirements
• avoid AP overlap between adjacent floors
• trim lower data rates and set local proble threshold
• 20MHz channels only

Even if I lowered the APs to 3dBM, the edge of the cell is still causing interference across 3/4 of the floor and to parts of the floors above/below.  I haven't worked with 224s yet and directional antennas, but suspect this may be the route to take to avoid high CCI.  I'm also carefully considering allowing some DFS channels.  Any thoughts for anyone that's had to contend with high CCI in a high density multi-floor building?

- How far apart are your access points?

- You are using 5ghz at 20mhz and you are still seeing high CCI when you drop the power to 3?  Again, how close are your access points to each other?  How many floors in this office building?

About 35ft between APs.

I've simulated 3dBm in ESS and it hardly made a difference across the floor.  I need to try it off hours on the equipment and do another survey to see if the results reflect the simulation.  Problem is, to meet coverage requirements, 15 dBm is the sweet spot.

Your lowest common denominator is the VOIP device.  If the voip device is a handheld device, that would be the impetus to start surveying  with APs at every 50 feet, but no less.  That would be because the voip device has a lower tx and rx.  If your VOIP device is a laptop, you should start at every 60 feet because they have greater receive sensitivity and tx power.  I'm not trying to throw shade on ESS, but predictive tools have their limits.

I would survey with APs at 12 power at 50 and at 60 feet and see what kind of propagation you get.  No matter what you put, you will get potentially 2 to three times the coverage due to above and below propagation.

Thanks for the input, Colin.

Just as you're saying, the Lync VRD suggests 50ft between APs.  I'm curious to know, where Aruba is getting that number from?  It sounds similar to the rule of 1 AP per XXXX sq ft, which I do not agree is the best way to approach wifi design.

thecompnerd,

Please read my exact phrase:

"I would survey with APs at 12 power at 50 and at 60 feet and see what kind of propagation you get."

I did not say Aruba, I said "I".  Also I did not say deploy, I said "survey".  These are my words and are from my experiences, and my experiences alone.  Every deployment is different, but you have to start somewhere.  That somewhere is with an actual survey and I would start it at 50 or 60 feet with the access point(s) at 12 to match the minimum power you would have an AP on the 5ghz band.

You have to make individual decisions based on your site specifics.  I am mentioning a starting point and my rationale.

The best way to approach wifi design is to start somewhere, prototype and iterate.

I understand that your views may differ from your employer, but you were stating something that sounded similar to the Aruba Lync VRD.  On the surface it appeared as though you were referencing the company's design recommendation and not your own.  I understand your rationale and appreciate the response.

You haven't addressed my question regarding the directional antennas, and I'm wondering what your thoughts are on that?  I thought I'd keep that option in my back pocket in case moving APs and changing power levels isn't helpful enough.

thecompnerd,

It is not about my views differing from my employer...it is about giving advice based on what you posted.

The ap-ant-1 (http://www.arubanetworks.com/pdf/products/ap-ant-1b_ss.pdf)  would provide more horizontal coverage with the AP-224, but is the difference in horizontal coverage enough to justify the complexity of ordering, deploying and maintaining a separate part when you can do the same thing with an AP225?  Would simply moving the access points further apart deal with your issues? THAT is what a survey will reveal.

With that being said, 35 feet is too close together because you would not be able to get any meaningful reuse at that distance in combination with a multi-floor deployment (your data already shows that).  Today's access points can transmit better and are more sensitive.  Start with coverage, based on your APs at a power of 12, THEN figure out if you have enough for roaming/density, etc..