Should you always do a site survey?

[dlee3370]

Currently we have an office that is about 5000 sq ft. Its divided into 2 half's so we bought 2 MR 42's that we have no problems yet (were not at max capacity). We bought out the rest of the floor which is an extra 7000 square ft. I plan on buying 3 AP. Do I need to get a site survey to determine where to put the new access points ? 

Thanks for your help.

[Uberseehandel]

Ask yourself, why wouldn't you do a site survey?

Not to do so could be misinterpreted, should there be any nasty surprises.

• known knowns
• known unknowns
• unknown unknowns

It is an opportunity to get to grips with Ekahau

 HeatMapper is free.

[BlakeRichardson]

My answer is yes, even if you do it yourself. Guessing it not the best option when it comes to working this out.

I am not sure if Meraki have partenerred with any software developers who make this sort of software but that would be an avenue or contact your local Meraki Rep to see if they have any suggestions. 

[PhilipDAth]

I am going to fly in the face of common advice, and say that I seldom do wireless surveys anymore.  Maybe down to 2 a year.

Typically I try to get customers to go for "good" or "saturation" levels of coverage.  I then place access points so they have a radius of coverage every 15 meters, or spaced 30 meters apart.

I suggest to many customers they are probably better off taking the money they were going to spend on us doing a site survey and simply putting that into more access points.

[MRCUR]

I totally agree with @PhilipDAth at this point. This is especially so with high density environments (like schools or hotels) where a survey is going to tell you to put an AP into every room to meet the client load requirements. Even in the opposite type of environment where there are a low number of clients, doing what Philip suggests is going to get you 95% of the way there. 

[Uberseehandel]

@MRCUR @PhilipDAth

I think you are both missing a point. 

A site survey shows up more than where to place APs, it show where not to place them, what problems there might be that will have to be addressed, and whether there are actual, or potential sources of RF interference that have to be taken into consideration.

[MerakiDave]

Ok team, I'll take the bait!  My answer is a solid "maybe".  Haha.  Simply because it's always an "it depends" answer.  Some will maintain that the first 3 rules are: Do a predictive site survey, do an active site survey, do a post-install site survey.  But it depends on your physical space, depends on what you're designing for (voice, location, HD, general coverage), depends on your applications (what's going to be driving the channel utilization), depends on user/client density, and it depends to a large degree on your WLAN design experience. 

Solid planning can often get you 90% home and AutoRF can do a great job to close the gap.  And manual settings can sometimes address any corner cases.  In a school with an MR42 per classroom design, or a hotel with an MR30H per guest room, or a stadium or large auditorium design, or typical carpeted cubicle office space with an AP per 2500 square feet... once you've done a couple of these deployments, it becomes routine enough that paying lots of professional services dollars for formal site surveys is not where you're likely to spend your IT budget. 

If it's a good-sized customer and they're not too experienced and they'll have lots of future deployments, I'll recommend professional services for an active site survey, as everyone should have to do that once!  I'm half joking, but that means I'm half serious too.  It's an excellent learning experience that forces a customer to understand (or at least appreciate) the design factors, interference sources, neighboring WLANs or other things they didn't know about, and from a design perspective it shows them that if they just go deploy APs haphazardly without proper planning, they're asking for trouble. 

In many other cases, and perhaps a LOT more often, I tend to recommend predictive site surveys, and there's plenty of tools and training to do so.  It's typically worth the little bit of time and effort to pop in some floorplans, set measurements and building materials, and start placing APs to confirm full and adequate coverage specific to your parameters.  Even if someone doesn't understand all the inputs to the design and the physics of RF, the resulting heat maps are simple to understand.  I also know and have customers who pay partners for less expensive predictive site surveys as a middle-ground on large-scale projects.

[EDIT - Afterthought] I forgot to mention there's a support article as well if you're new to this and wanted to leverage a Meraki AP in Site Survey mode.

https://documentation.meraki.com/MR/WiFi_Basics_and_Best_Practices/Conducting_Site_Surveys_with_MR_A...

[thewifigeek]

It depends on the situation.  Always have a purpose/intent of why you are doing it. 

In your scenario, no you dont need a pre-implementation survey and place the up to 3 APs in common sense areas away from obstructive/reflective surfaces and ceiling mounted.  However if you're doing a design for new customer then a pre-survey is a great opportunity to also audit the environment for realistic mounting positions, onsite spectrum analysis, photos of switches (port capacity, power availability), and good old fashioned meet and greet face-to-face.  Other reasons:

1.  simulated design is to provide report for the cabling team for installation

2. to visually showcase the proposed design will solve their coverage/capacity problems.

3. etc.

If you're designing a harsher environment (e.g. Hospitals, outdoor WIFI) then 100% you simulated and test propagation onsite with actual access points on tripod.

I encourage post-implementation surveys to validate your own work (against a simulated design) and the report will provide guidance to operational support teams which is often not yourself "RF engineer".  Also during the post-survey I encourage 802.11 analysis so you can optimise contention levels.  

The challenge with surveys is that they take time and cost $.  So depending on the situation I would encourage surveys as it provides level of assurance and removes some guess-work which is the difference in solving a WIFI issue in 1 month (trial-and-error remotely) or 1 week.

Hope you find this useful

thewifigeek

[Uberseehandel]

apart from anything else: if the people responsible for WiFi carry out a survey using a tool such as HeatMapper (conveniently free), they will learn more about the RF environment in which the network functions, and may even add to their own understanding of how WiFi functions.

Living in the Old World, buildings tend to be more complicated constructions than New World constructions (one of my buddies, an SV megastar comes over to Europe and wonders, aloud, why people build houses to last more than 40 years).

Even in a moderately solid building with brick interior walls, I register upwards of 25 adjacent WiFi systems. One of the uses of a site survey is to provide some quantifiable information upon which to base a decision not to use the 2.4 GHz portion of the spectrum in most multi-tenancy situations. 

Both my lab set-up in the UK and the design/testing bureau in France are adjacent to marine and aviation operational facilities. There is still a great deal of equipment in use that triggers DFS/TPC/CAC systems to kick in. Another of my colleagues has a site adjacent to a busy airport 290 km N.E. of my location. When storms roll in from the Atlantic, I warn him that the APs at this site will experience outages (CAC) and that the channels in use will change (DFS).  This doesn't occur because of ILS, but rather because one of the runways is so far away from the terminal that some of the elderly vehicles used by security/safety/guidance staff rely on out of date RF equipment to be sure of not getting lost . . . (funny but true).

Even on domestic deployments it is important to understand what is involved. At one stage foil backed wall board was popular (even in Aotearoa), the varieties I've come across function like a Faraday cage.

[dlee3370]

Thanks for the great replies. I think I will go ahead and get a site survey done. It will help me learn the process which I can use in the future. 

Thanks again!

[Uberseehandel]

@dlee3370

I really hope you enjoy carrying out the site survey. There is a great deal of really helpful documentation available on the Ekahau web site.

[thewifigeek]

Best of luck and enjoy the journey!

It's very rewarding when WIFI issues essentially tell you their problems.  This comes from years of asking "why".