Today there has been an argument between someone, who meant to be a so called „pro“ in onsite surveys and me. We found out, that he didn’t know even the basics. It came to my mind to do a short write-up on how usefull onsite surveys should succeed by using todays technology.
This is an introduction to onsite surveys and nothing more. There’s a lot more you should be aware of – I assume most of you already are from a technical perspective. This writeup is on a very basic level for others such as customers don’t know how onsite surveys work in general.
If you do follow me here over or on some forums (mostly german language), you might assume I might know a lot on how wireless design should be, but that isn’t true. There are people like Adrian Granados or Keith Parsons I’ll take ‚em up on almost everything they write or talk. There’s still so much on WiFi I have to learn and do so every single day.
Yes, there has been a time about 10 years ago I struggled a lot and didn’t know why WiFi didn’t work exactly as I expected. I couldn’t predict or sucessfully architect before I did at least some deep dives. I had to learn a lot about WiFi’s kind of „organic“ behavior. When it comes to very high density (VHD) environments, all of these elements I didn’t know aggravated.
Today my initial situation differs, as I can use all the tools I need and I do have at least „knowledge“ I’d also like to share with you what’s most necessary.
First: It doesn’t matter what software you use or to what manufacturer you design for, there are always the same necessities – even if every single customer demand as well as every single building and every single network purpose differs from all the others – the content of this article.
WiFi is no switch!
1. Talk to customer, count the numbers!
As you know that every single onsite survey differs, you don’t know customer demands. As we no longer plan on coverage, you will need to know, what purpose your WiFi architecting is for and how many client devices or what application will use this network for airtime/capacity calculation.
This is often hard work as you don’t get the exact information easily by some customers. Sometimes it means a little more than just talking to key users or the boss. It is a mixup of customers explanations, corrections and assumptions. You’ll need at least total numbers of client devices, device types and applications they might use in a worst case scenario. Also some demands cannot coexist to others – so you’ll need to talk through that and explain.
I’d second – the better your results are, the more people will use customer’s WiFi in future. Keep in mind planning even more dense including higher numbers by decision of customer.
In addition: If you do have those numbers, tell them, how much bandwidth is necessary and calculate at least 50% on top. If customer’s carrier can’t provide adequate performance, your airtime calculation won’t succeed. You can ask them if they will do survey with you next year after they upgraded. They also should be made aware of throttling is the worst ever they can do to WiFi but I see all over this planet in most installations because of people don’t understand WiFi.
There are also decisions to be made, e.g. for school districts: Are students allowed to use BYOD or teachers only? Give them information also on TCO by explaining the difference BYOD yes and no in planning and providing.
Summarized: Get to know exactly what they want and tell them what WiFi can do for them within or right after interview. The very key element is: Keep everything on paper (as you made assumptions) and ask your customer by email if your writeup is all-right!
2. Grab the plans but have a look onsite also before simulation!
Plans don’t mean anything most of the time. For me a plan is nothing more or less than just the basis I’ll draw my own walls on. Either they’re not accurate or you really don’t know what dB that special wall is. Maybe you’ll bring your AP-On-A-Stick and your Sidekick with you to get those dB numbers before you start Simulation. You can easily combine this with your customer interview you’ll have to do anyway.
3. Simulate ON PURPOSE!
There are lots of informations you gathered by interview already. You also agreed on a manufacturer and series of AP’s you can provide. Any of those bigger manufacturers do provide implementation guidelines you should be aware of. If you plan Aruba in a big school district, you should know the Aruba VHD design guideline (Link) for example. By relying on those guidelines, you’ll be safe.
By choosing your onsite software, your tool should be supporting exactly those AP’s and these guidelines. You should also be aware of mounting positions. When it comes to ac or future ax networks, this is even more essential. To gain airtime for your devices (e.g. to reach MCS9 if possible), you will keep away anything that influences line of sight to client devices as much as possible. There are several options you can choose from by e.g. oberon mounts (Link). This also enables features like „beamforming“ to lower interference and it is necessary for location based services.
As you got to know the numbers in interviews, you can now use your software and plan exactly for your customers needs. The customer must also know if customer wish is difficult or impossible to implement.
If I get results of an onsite survey that is all „green“, I don’t trust and run away. There always is an area that doesn’t fit to your guideline and customer must know beforehand. In most situations it is impossible to have it „all green“ based on a manufacturer’s guideline.
You can and mostly should also segment your planning on customer purposes and numbers. The easiest example is if you segment an office to a meeting room to a school’s auditorium.
By drawing walls placing simulated AP’s you should always keep in mind the following key elements (in no particular order):
- coverage
- interferences (all types as they primarely define performance)
- capacity and purpose/applications
- secondary coverage
- signal to noise ratio (that is your highway you’ll be surfing on)
Let me point out something very clear: If one or some of those above listed five elements are unclear to you, you shouldn’t start designing. Read a book instead.
In addition: Keep in mind motion zones as they highly affect dedicated zones, especially while designing BYOD. In Germany we do call floor mounted AP’s „Fliegenfänger“ (flypaper) and they’re necessary (outdoors IP68 also!).
4. Do onsite passive surveys with suitable tools and do it again after implementation
You will count on working equipment. You must know how your gear behaves. You’ll need those AP’s to demo situation customer will use later on. Switch off all the automatics such as ARM and configure standard dB ratio as well as fix both 2G and 5G radio channels on all your test-AP(s).
Yes, I know of portable battery solutions are very expensive but there’s nothing worse than you being onsite and your DIY-solution doesn’t work. So this should be a well planned investment. As I know of my time is money, I can cut in half costs on the long term by doubling my test AP’s on a stick. Triple them if possible or your boss allows you to do so. Make him aware of moving and realigning test AP’s for the next survey-round is the most time consuming element.
As survey people do like MacBooks (they come with 3SS Radio like mostly no other), they do come with one of them only and you won’t use them. There are now solutions on the market, you should rely on if you want to have reliable and quick results, like Ekahau Sidekick (Link) for example (I love it). First of all they do have internal batteries that don’t suck your Surface-Tablet battery and second they do have reliable and so quick high density radio modules, you can have your point-and-stop or even on-the-go measurements in a fraction of time than using those standard USB-Sticks on a on-laptop-glued hub that cracks in pieces while bumping with something against. You even could do surveys on a Segway by using those solutions.
The result you gathered in your software tells you if your simulation is good or it simply isn’t. Most of the time, my drawed walls differ to measurement and I will re-simulate based on measurement results. I am now aware of any signal interferences by third parties and I know the noise floors by other elements such like microwave ovens, generators, DECT-cells, household WiFi-cells, LTE and others unusual elements that pester airtime and lower my SNR.
Recalculate your simulation!
After measurement your software can recommend channel and dB configurations on any AP for that scenario including external interference and your co channel interference to meet your customers and guideline needs. You will add, remove or replace AP positions. Over and over. And yes, I know, this is „a game like Tetris“. You can compare this to loading your Vauxhall Astra for holiday journey having three kids.
5. Hand out and talk!
I found out, there’s maybe a better way to hand out your result to customer. Technical details of your measurement do not interest decision makers but are necessary to support your results. Your onsite survey results will therefore be divided into two parts:
First part
will contain scenario description, worst case numbers that you found out in interview, explanations, part numbers and counts, the positions your AP will be mounted to (copied from part two) and a suitable list of AP’s including configuration (you can also leave a blank field next to AP name/number, the electrician will stick the mac address onto, to hand it back to you for revision or controller configuration).
Second part
will contain the raw report your software generated that lead to first part
There are several people that will speak up against, for example electricians or people like fire protection officers. You maybe have to adjust your recommendations and talk again.
After implementation re-survey, tweak (repeat) your AP’s controller to fit needs.
As you know I do prefer Ekahau. There are lots of competitors that work also and provide almost the same features. You need to test yourself. This is just how I do it. Maybe you’ll find better solutions or do have better suggestions, so please leave me a short note by email or twitter.
John