Technology Blog

Ultimate Guide to Modernizing Classroom Wi-Fi

by Moderator on ‎11-24-2014 05:58 AM - last edited 3 weeks ago by

New digital learning tools boost student success and empower the next-generation of teachers. But they are only as good as the Wi-Fi infrastructure they run on.

 

Aruba recently hosted a webinar with industry expert Keith Parsons who explains how to create the perfect Wi-Fi network for one-to-one learning, how many Wi-Fi access points you need in a classroom, and how to choose the right Wi-Fi technologies to meet your needs.

 

Keith Parsons, Managing Director, CWNE#3 Wireless LAN Professionals, is a gifted presenter with extensive experience designing and installing Wi-Fi in over 2,000 K-12 classrooms.

 

View the webinar slides and Q&A below. For detailed explanation, check out the on-demand webinar.

 

 

 

 

Q: As a shared resource, slow devices take proportionately more time to send/receive data than a faster device. Then should we be creating a separate SSID for very slow devices (like iPhones, etc.)??

Slow devices do take up a proportionately higher amount of the available bandwidth on any given frequency. Moving them to a separate SSID does absolutely nothing however. All SSIDs on the same frequency share the limited resource of time. SSIDs do NOT differentiate in any way in the airtime. Though you can tie an SSID to a VLAN, they do NOT work at all like VLANs on a switch.

 

Q: Any recommendations on SSID's for K-12 education. One vs. many?

Usual recommendation is the fewer the better. Each additional SSID takes away from your available capacity. Try to use a new SSID only when there is a different authentication type. You can also look to having a Dynamic VLAN assignments based on RADIUS attributes to save even more SSIDs.

 

Q: Any recommendation on design software for Wireless LAN Access Point layouts?

  • AirMagnet Survey Pro
  • Ekahaus ESS
  • Tamosoft Tamograph

 

Q: By "4 radio coverage", do you mean 2 SSID

No. I mean two radios covering the area in 2.4GHz and an additional two radios covering the same area in 5GHz. Basically two Access Points covering every area at -67dBm will meet most of your capacity requirements. Even with this you most likely will have to turn off a big chunk of your 2.4GHz radios to cut down on Co-Channel Interference.

 

Q: Can you recommend tools for CCI detection?

Any client device that can show you Access Points and their BSSID’s (MAC Addresses) and Signal Strength will work. (inSSIDer app or the Aruba Utilities Apps as well as most client device software)

 

What you are looking for is two or more access points on the same channel with signal strengths (RSSI) above -85dBm.

 

Q: Can you set APs to block sites like YouTube for a group of users?

Yes – it depends on your vendor of choice. But this capability is built in to enterprise-class vendor products.

 

Q: Do you recommend any particular Wi-Fi mapping/testing device and/or software?

I like a Fluke AirCheck for lot of my on-site testing. For validation surveys you can try:

  • Fluke AirMagnet Survey Pro
  • Ekahau ESS
  • Tamosoft Tamograph

 

Q: Do you recommend turning off the "B" channel or 2.4 altogether?

Definitely turn off the 802.11b data rates. (1, 2, 5.5 and 11Mbps). You’ll probably still need to leave on 2.4GHz for legacy clients you are emotionally tied to. (see Sunk Cost Fallacy). But most likely you’ll have to turn off many of the 2.4GHz radios in order to reduce CCI to a workable level.

 

Q: Does an Aruba 3400 controller handle frequency reuse?

It doesn’t. Not part of a controller’s function. The Auto-RF algorithms will try to lower Tx power, but don’t have the ability currently to totally turn off radios when someone puts in too many access points.

 

Q: Does creating a "Guest Network" on a separate SSID alleviate the problem of older devices slowing down my Wi-Fi?

SSID’s in Wi-Fi do NOT change collision domains whatsoever. (they also don’t change Broadcast domains either) – on the Wireless side of an AP, all transmission share the same amount of available bandwidth – changing SSIDs does nothing to fix a slower device, or CCI.

 

Q: Doesn't Aruba's "mesh" and "ARM" technology mitigate CCI for AP's that are densely deployed?

Neither. Mesh replaces an AP’s wired backhaul with a wireless uplink to another AP with a wired connection.

ARM tries to adjust power and channel to minimize CCI – but cannot overcome a situation where there are too many AP’s because of a poorly engineered solution.

 

Q: If a school was designed with 5ghz access points, when upgrading will the Access Points will design and placement of the access points need to be redesigned?

I suggest each time you do an infrastructure upgrade you return back and do the Define, Design, Deploy and Validate process once again. You may have in your Define stage a constraint about where current AP cable drops are – and this can be used as part of the Design process.

 

Q: 802.11ac future specs?

Right now the 802.11ac specs are completed and you can read them on the IEEE website. As for Wave 2 – most of those are also available but as working group documentation.

 

Q: Cell -db overlap setting on our survey tools?

Try reading the ‘Fallacy of Channel Overlap’ white paper on my website: http://www.wlanpros.com/wp-content/uploads/2010/06/The-fallacy-of-channel-overlap1.pdf

 

Q: Should schools be running 2 x Cat6a drops to classrooms now if they have the funds? (refresh cycle giving us this hint)

A lot of the costs of running cable is in the labor. Adding an extra strand of cable to each AP drop won’t cost very much additional… but might pay off handsomely in the future. It is a prudent thing to do if you are refreshing your cable plant.

 

Q: We usually filter our Surveys for cell overlaps at -76db as a "roaming" point. Is this a good plan?

If I understand your statement – I would strongly disagree. Setting your cell overlap during a survey at -67Bm means you never walk and collect data in the areas outside -67dBm… this is where most, if not all, of your CCI problems will exist. So if you don’t collect data there, you’ll have an overly optimistic view of your WLAN. Then end up with CCI issues and capacity reductions because you never measured it.

 

Q: How do I figure out if I should use a 20 MHz channel or 40 MHz channel?

Always use 20MHz in 2.4GHz. I like using 20MHz in 5GHz as well. It allows for twice as many channels and thus easier frequency reuse. Until the majority of your clients support the 40MHz channels AND support DFS channels, stick with 20MHz in 5GHz as well.

 

Q: How do we know we've found a vendor that will help us engineer the best solution?

By asking lots of questions, ask for references, and be comfortable with their ENGINEERING skills, not just their marketing techniques. Ask about their engineering design processes – and confirm they are doing a good job of working with you to define your detailed requirements, and using appropriate engineering design techniques and tools, and then always ask for a post-installation validation survey.

 

Q: How do you match the capability of your Wi-Fi to the broadband capability the telco is trying to sell? In other words, does moving from 20 Mbps symmetrical to 100 Mbps.

Two different, yet related items. You might have a fantastic Wi-Fi network, with a terrible WAN connection. That would look like a wireless problem to your staff and students. Having a fast Internet pipe, and terrible Wi-Fi would exhibit the same complaints. But the solutions are different.

 

Q: How long will it be when 802.11ac Wi-Fi standard going to be found in over 50% of the devices?

Already all Smartphones available today and in the future have 802.11 chips on board. You can help this buy purchasing exclusively only gear with 802.11ac on board. BYOD will probably also all be 802.11ac capable from now on.

 

Q: How much radio planning is required to be done manually compared to how much is done automatically by the controller?

RF planning is not done at all in the controller. Only Tx power and channel choices. The actual placement of access points and use of walls to attenuate the signals to help with frequency reuse is a physical world issue.

 

Q: I work in a school district that is doing a 1:1 distribution of iPads. We have run into issues with excessive amounts of bonjour traffic. What are your suggestions on how to manage that?

First, be sure all your Apple TVs are wired and not using Wi-Fi for network access. This will help a lot with your wireless network throughput loads. Next – use your vendor’s solutions for Bonjour ‘gateway’ services to keep Bonjour’s broadcast traffic contained.

 

Q: Is a spectrum analyzer worth the investment?

Yes – today you can get SpecAns for $500 to $4000 in the Professional ranges designed specifically for Wireless LAN usage. They are very helpful in seeing the raw ambient RF interfering devices.

 

Q: Is front end overload from nearby APs that are off-channel a concern? Beyond adjacent channel interference? This question may come down to "how good are the filters?"

You should never place adjacent Access Points within a meter of each other. Their front end RF processers are not designed to handle the extremely high power levels within the 1 meter range. Also – you should design your channel plans to NOT have adjacent channel interference in the first place.

 

Q: With BYOD and 1-1 this means 18 to 30 devices per room and so if there is a limit to the number of devices using an AP - coverage and Density are two different things... right?

Coverage is easy… density a bit more difficult. What is very difficult is getting rid of Co-Channel Interference and designing IN frequency reuse. It is the only way to really get the extra capacity we need for the levels of clients you are talking about.

 

Q: It's easy to say don't get 2.4 Ghz devices on your network. Unfortunately, many laptop and computer manufacturers still are making 2.4 GHz only devices. This basically means that the computer and hardware manufacturers need to also get on board.

But you don’t have to buy them! They produce only what people ask for and purchase. You can ask your purchasing department to “Never buy 2.4GHz-only devices again.”

 

Q: Can you still make it work with both 5 and 2.4 since this is still an issue?

Usually you’ll end up with a 5GHz plan that works well – and then you’ll be turning off ½ or more of the 2.4GHz radios.

 

Q: Max throughput for 802.11ac in the future? Any guesses? We are running two cables to each AP. Waste of resources?

Nope – I think running two cables to each AP location is a great use of resources. Max throughput… hmmm… perhaps very close to the gig Ethernet of today – but also could be higher. Thus bringing along a spare Ethernet drop might be needed.

 

Q: On average, how often should APs be replaced?

I tell my customers to think 2-3 years less than your wired infrastructure. Perhaps instead of a wired infrastructure update every 5-7 years – think and plan for Wi-Fi in the 3-5 year range.

 

Q: Planning Phase? Is it possible to design for client flexibility? Today it's iPads tomorrow its ChromeBooks, or Win8.1 Tablets, or product X. How do we account for a future that is unknown.

That is why you spend a lot of time in the Define phase. Gathering not only requirements for today, but making educated guesses about the future as well.

 

Q: Refresh cycle? Is there a need to redesign for all hardware refreshes? Or only if bigger changes? Example, redesign from 802.11a to 902.11ac?

I’d recommend a complete process including Define, Design, Deploy and Validate every time you change the base infrastructure. (from .11n to .11ac) You can add back in the constraints for the design phase by recommending AP cable drops only move when necessary.

 

Q: So let's say you're one of those schools with Aruba 225's in every classroom, and the performance isn't sub par. What do you do now? Is a spectrum analyzer the answer?

You might not know if your performance is sub-par. The downside of too much co-channel interference doesn’t raise its ugly head until under heavy loads. You perhaps haven’t stressed your Wireless LAN to that point yet. Nothing at all wrong with having One AP per Classroom… IF that is the result of properly engineering your solution to meet your requirements. I have just found many don’t do post-install validation surveys, never measure CCI, and don’t load test. Spectrum Analyzers won’t do anything if you have too many AP’s all sharing the same frequencies. The answer is to turn off many of your 2.4GHz radios, and perhaps even some of the 5GHz radios if you are measuring co-channel interference there as well.

 

Q: How often should an installed solution be reviewed in depth?

At least every time there is a significant change in the infrastructure. Changing from 802.11a to 802.11ac for example. Or when you have very large changes in client devices. Like adding 1:1 iPads for all students.

 

Q: How do you feel about using the UNII (2) channels in regards to channels getting shifted due to radar.

I love using DFS channels. It makes getting frequency reuse that much easier. But you have to be careful and confirm your target client devices can see and use those channels. If not, you are designing in coverage holes for those devices. In that case, drop back to only the channels your 5GHz devices can use.

 

Q: Understanding that it not a good design strategy is 1 AP to 1 Classroom ever the right solution?

Of course it is! If after defining your requirements, the design calls for that density of access points, and you do a post-install validation that confirms you don’t have co-channel interference – then of course you have a design that has been engineered to meet all your requirements. If, on the other hand, you don’t do a validation survey, don’t measure where you have CCI. Then you might have extra AP’s that aren’t adding any additional capacity. That would mean you spent more than you needed to.

 

Q: What frequency scanning tools would you suggest to best do the "first glance" to get an idea of how to design/improve a wireless installation?

The fastest, quickest way is to use a validation survey tool to confirm where and what your RF is doing in your facility. You can also use this to prove and verify if you are meeting your design goals and see where you may or may not have CCI issues.

 

Q: What the most practical way to eliminate CCI? Or the proper term CCC.

Use properly engineered designs that allow for frequency reuse rather than cause CCC. Usually it involves doing an on-site post-validation survey, reviewing the results, and then turning off radios – moving AP’s and perhaps even adding directional antennas where appropriate.

 

Q: Where can I learn more about CCI?

I always recommend starting at the beginning with something like the CWNA Study Guide by Coleman and Westcott. http://www.amazon.com/CWNA-Certified-Wireless-Administrator-Official/dp/1118893700/ref=sr_1_1?ie=UTF8&qid=1416359756&sr=8-1&keywords=cwna+study+guide&pebp=1416359757905

 

Q: With a so complex WLAN engineering design, how can you make a predictable price offer to customers?

By doing the engineering properly, it isn’t merely a ‘prediction’ but you can get very close to the actual pricing. The RFP should include the engineering first in order to even know what the rest of the RFP should include.

 

Q: Would you recommend using the auto channel selection feature of access points and letting them decide their own wireless channels vs manually picking the channels?

This is a scale issue. In smaller facilities I like to do the channel plans manually so I can control exactly what is going on. But as you scale to bigger facilities, this gets a bit cumbersome and time intensive. So the vendors have produced algorithm-based wizards that help your access points move to channels and stay out of each other’s coverage zones as much as possible. I still like reviewing manually what the controller does. Any post-validation survey can be used to see whether or not it met the RF requirements. But remember, algorithms can only work with in the installed AP’s and can’t turn off individual radios to reduce CCI.

 

Comments

Really good article, just one query:-

 

Q: Does an Aruba 3400 controller handle frequency reuse?

It doesn’t. Not part of a controller’s function. The Auto-RF algorithms will try to lower Tx power, but don’t have the ability currently to totally turn off radios when someone puts in too many access points.

 

Aruba's kit does have the "mode aware" feature which enables ARM to turn off 2.4 radios if it feels intereference levels are too high. I'd argue that manual selection of these is probably safest.

Actually, one other thought...

 

I give a massive thumbs up to this post as a good "one page how to" document on how to design WLANs in general - The general practices that Kieth recommends apply to many, many situations, especially where high capacity is required. It's really good to see Aruba engaging with RF desgin in a serious manner.

 

This is not, however, what is preached across the board. Whilst Left Hand of Aruba is saying these good things, Right Hand of Aruba is saying something completely different. I recently attended the 'AWW 801.11ac Migration Best Practices Webinar'. It was quite high level, but really gave the impression that one could just do by-default "1 for 1" replacements and stuff like site surveys really is optional / only applicable in environments like warehouses. 

Search Airheads
Showing results for 
Search instead for 
Did you mean: 
Announcements
Read all about it! If it’s happening now, it’s in the community.

Check out the latest blogs from your community team, the community experts and other industry sources.
Labels