40HD Independent wireless streams in one building… a world first?

Ramblings of an IT Guy in Education

40HD Independent wireless streams in one building… a world first?

September 6, 2018 Education IT Infrastructure Ubiquiti Wireless Video Presenting 0

I remember when I graduated from my high school in 2003 I thought it was pretty cutting edge that each of their computer rooms had a projector. Jump forward another year I was a young IT Trainee in my first school, it was an independent Catholic school. They had 6 projectors and not just in computer rooms but in some of their general rooms. All the other rooms had mostly whiteboards, and in the next school, I worked as a System Administrator, they had more projectors, but not every room though they were cutting edge enough to not have any blackboards.

Why am I telling you about how the room looked almost 15 years ago… to illustrate the difference in classrooms over time.

Now in 2018, it is mostly standard that every room in a school must have a projector and wireless. For the last 4 years, a lot of schools have been attempting to tackle the next big IT Change in the classroom. Wireless Projector. To remove that tether from the front of the room, or to completely remove a front of the room.

In Term 1 2018 CCGS opened the doors of their newly renovated B Block. With teachers as the driving force behind the designs, a design that is capable of 40 HD independent wireless streams in one building. From what we can tell it is a world first. And it has completely changed how teachers use the classroom.

NOTE: This is a long one, hence why I have made sub-sections. Please see links below if you would like to skip to a certain section. You will also see WVP in this post, that stands for Wireless Video Presenting.

Why is WVP so hard
Is my wireless strong enough for WVP?
Basic WVP
HD WVP – Source mirroring
HD Video WVP 1:4
HD Video WVP 4:4… and beyond
Products we tested
The Initial design
The Pilot
The Wishlist
The Mesh… bring on the Faraday cage
Summary benefits of HP WVP
Teacher and Engagement
Crucial Aspects of doing WVP well
Vivi Agent… quick run-through
Conclusion

Who was involved?

Before we dive into this long post, I feel it is important to acknowledge those who were directly involved in this process. They are more people involved who were part of planning committees that brought in recommendations or a list of what they wanted in the rooms. Though I am really focusing on the people who brought this from a hand-drawn design on a blank piece of paper to the achievement it has become.

David Soede; Director of ICT

Damon Cooper; Director of Teaching and Learning

Ben Collier, AV Technician

Kyle Giannini, Helpdesk Technician

Why is WVP so hard?

For many years, schools have been trying to tackle the beast that is WVP. However, it is something many years schools have tried to succeed at, but try as they might it is rare to find a school that has  completely replaced the need or the desire to use an HDMI cable. Because unlike HDMI cables you need to think about all of the following things and ask questions about each one.

  • Source content quality; where is your
  • Hardware device quality
  • Operating system & drivers
  • Playback software & codec
  • Wireless performance
  • Wired network performance
  • WVP Platform hardware & firmware
  • Display device quality

Let’s move onto the wireless network because when you want to broadcast a wireless High Definition image, that is one of the most important parts.

Is my wireless strong enough for WVP?

We have all seen the online ads or emails. Wireless Company has an AP capable of achieving an ungodly amount of bandwidth. Though it is normally marketing bullshit, those results achieved in unrealistic setups, like a faraday cage deep in an underground bunker where no possible chance of interference exists.

I rolled out Ubiquiti UniFi AC EDU in each classroom at Christmas, each one claims 450 Mbps on 2.4 GHz and 1300 Mbps on 5 GHz. A total of 1750 Mbps, which is again complete bullshit. Firstly, I don’t run the 2.4 GHz range in classrooms, there are just not enough channels for me to be able to reliably run it. So, it gets turned off, and now that speed is down to 1300 Mbps. Then I only use 1 of the 2-gigabit ports (you can’t use both without ugly cabling, design fail or Ubiquiti know you won’t need it?) So Best I can achieve then is 1000 Mbps? Well no, it doesn’t work like that. And this isn’t Ubiquiti, Aruba, Cisco or any other wireless vendors fault. It is just how the technology works.

If you look into the breakdown below and you can see what you get with a 40 Mhz channel vs a 20 MHz channel and the number of spatial streams, you get much less. If you look at the graph below (Figure 1) you can see that you will get between 300 to 400 Mbps… if everything is perfect. You can potentially get much less. To achieve that magical 1300 Mbps you need to be using 160 Mhz channels which in almost all cases is just not possible. Especially with 150 AP’s on my campus.

Figure 1. From https://en.wikipedia.org/wiki/IEEE_802.11ac#Data_rates_and_speed

You can read more about bandwidth breakdown and what you can really get out of your wireless on this site, it is a fantastic read: http://araknisnetworks.com/2017/06/dude-wheres-my-throughput/.

We had chosen Vivi as a WVP platform, and on their Admin Guide (https://www.vivi.io/hubfs/User%20Guides/Vivi_Admin%20Guide.pdf) it tells you that you need 16-100 Mbps of bandwidth for each Vivi. This is basically the difference between a talking head or PowerPoint vs something really fast moving. Like an action scene, think Star Wars type action, fast moving lots of things happening on the screen.

So if we take the best possible wireless speed I mentioned above of 400 Mbps, let’s do some maths around that in table 1.

1 WVP pre-room Design
AP Bandwidth 400 Mbps
Vivi (1 stream) 100 Mbps (up to)
Remaining 300 Mbps for up to 30 clients.

Table 1.

Table 1 shows that is is very achievable with 1 AP and 1 WVP in the room, and I was more than happy to sign off on a solution. However, want if you wanted more than 1 in a room… let’s say 4 Vivi connected to 4 screens.

4 WVP pre-room Design
AP Bandwidth 400 Mbps
Vivi (1 stream) 100 Mbps (up to)
Vivi (1 stream) 100 Mbps (up to)
Vivi (1 stream) 100 Mbps (up to)
Vivi (1 stream) 100 Mbps (up to)
Remaining 0 Mbps for up to 30 clients.

Table 2.

As you can see in Table 2, we were going to run into some series problems. Now yes, that 100 Mbps is the maximum that a Vivi may run at, and in real-world classroom situations we may never see anywhere near that. But as an IT Professional, I couldn’t risk that we might ground the room’s network to a halt, it would defeat the purpose of what we were trying to achieve.

As a result, we came to the conclusion that we would need multiple access points in a room. Though this isn’t a simple fix either, because it is all fine and good to add two access points to a room. But how do you make sure all four Vivi steams aren’t on one AP. Remember 400Mbps is the best possible speed, that number can drop. You can read more about how I went through that here; (https://www.eduitguy.com/2018/04/29/part-2-testing-wireless-broadcasting-and-load-balancing-based-on-bandwidth/).

 

Basic WVP

Achieving basic WVP is pretty easy. We are talking about a OneNote file or a website. This type of WVP is very low bandwidth intensive, and even if you experience issues with laggy slow screen refreshes the users won’t really notice, you’re talking about static screens.

Figure 2. Basic WVP.

HD WVP – Source mirroring

This is where it starts to get interesting. HP Video generally requires low latency and potentially up to 100 Mbps of bandwidth because it is uncompressed, unlike 1080p MP4 files which generally only need around 5 Mbps.

This is when you are mirroring your screen wirelessly while playing video. There is two type of videos that you will be showing, a talking head (interviews etc) or action-based ones such as playing Star Wars. The talking head type video will not be much of an issue. It won’t hit anywhere near your 100 Mbps limit, but an action movie can hit that limit.

Figure 3. HD WVP, mirroring your source.

So in this scenario talking heads would be fine with a single AP whereas anything more than that would cause bandwidth and latency issues which in turn will bring about pixilation, artefacts and stutter depending on the source of your file. Both Basic and HD WVP can be easily achieved with something like an Apple TV or Chromecast, a 1:1 mirror of your screen over wireless.

Even with this scenario because you are doing 1:1 the load on your AP should be okay, and to date, we have found that in rooms with 1 screen 1 AP at 40Mhz channels is okay.

HD Video WVP 1:4

This is where things get harder. It is at this point it can go 2 ways.

Consumer grade; this is using something like Apple TV’s or Chromecasts, each screen you connect to is another wireless stream, so connecting to 4 in a room is actually going to take up lots of bandwidth.

Industry grade; for a product like Vivi, when using 4 of their units in 1 room you don’t actually send 4 individual streams off to connect to all four screens. They actually use your wired network (assuming the Vivi’s are connected via ethernet) to broadcast the image to the four screens, see Figure 4.

Figure 4. 1:4 WVP with Vivi and Ubiquiti

The benefit of this industry grade type of WVP is that you don’t have to stop at 1:4, you could achieve 1:100 providing your networking backbone is strong enough. However, the impact on your wireless would only ever be 100 Mbps max.

HD Video WVP 4:4… and beyond

The hardest setup to achieve. As you can see in Figure 5 you have 4 users all sending their own HD independent stream to a screen in one room. To achieve this you need multiple Access Points to handle the load. However, once you have the infrastructure in place it allows you to go.

 

Figure 5. 4:4 WVP with Vivi and Ubiquiti

And you could grow this for example to the below example, and in theory, you could keep expanding, though I will be honest we haven’t tried above 8:8 in one room. We have however seen well above this number in one building across multiple rooms and classes.

Figure 6. 8:8 WVP with Vivi and Ubiquiti

Products we tested

A quick google will show you a huge amount of WVP products out there, and a lot of them claim to be the best solution out there. So how do you get the right one, testing? Ben Collier and I dived into the myriad of options that are out there.

Just some of the products we reviewed…

  • Vivi
  • Apple TV + AirParrot (for windows devices)
  • Creston AirMedia
  • WiDi
  • Kramer
  • Chromecast
  • ActionTec
  • AirTame

After our extensive testing, we came down to two possible options.

  1. Apple TV
  2. Vivi

At first Apple TV was the winner, when running test YouTube wireless screen mirror (4 at once) we were getting the best possible picture. At this time Vivi was on version 1.5 and we found it had promise but just wasn’t good enough for what we wanted to achieve. We also found that their client was not great, for example, they have a YouTube handoff feature. This is great because you can play YouTube videos on there without having to play it wirelessly and use up your bandwidth, but the process to do this handoff was too many steps and I knew that staff wouldn’t do it.

Although when version 1.6 came out everything changed, their WVP performance was better than the Apple TV’s, and they YouTube handoff was amazing and down to 3 steps, something I knew the staff would do. Once version 1.6 came out we went from flavouring Apple TV’s to Vivi’s, which I love because the idea of supporting Apple TV’s on a corporate network gave me nightmares.

So now we have a network solution picked, we also have the WVP product selected, time to move onto what we would do in B Block.

The Initial Design

In 2016 Damon Cooper (Director of Teaching and Learning) brought an idea to David Soede (Director of ICT). He wanted to change the classroom, he had an idea for some breakout spaces located in the classrooms that would allow for collaborative learning.

Drawn on a piece of paper, he wanted a screen in front of a desk with a whiteboard to one side where the students could sit/stand around it can take notes on the board.

Figure 7 – The Initial Design

The Pilot

The initial design was done on the cheap, the ICT Department sourced a cheap Kogan TV to be used with a single HDMI cable hanging down (We didn’t even try WVP). A small desk was placed in the corner and we found a spare whiteboard to mount next to it.

Simple but effective, a few of these pilots were set up and we found that the students loved working in small groups. It wasn’t just the gimmick; the students were able to collaborative learn in groups. And the teachers were able to plan their lesson not just around having 1 projector in the room but having a breakout space.

These spaces became known as Digital Colleberation Zones or DCZ.

Figure 8 – The pilot

The Wishlist

In 2017 after CCGS had run the pilot program for the DCZ, and the decision had been made to renovate our aging B Block it was decided it would be the perfect place and time to expand on the pilot.

And I couldn’t have been happier when I heard that the initial discussions around how the classrooms would be designed had left the ICT Department out of it. Don’t get me wrong, I am not saying the ICT Department should be left out of building work, far from it. The ICT Department should be involved in all building work since so many new and renovated buildings require IT components. What I did like was that while the teachers were working out what they would like in the classroom or more importantly how they would like to teach in the classroom should exclude the ICT Department, at least at first.

You see, I have found that IT people like to manage expectations, which is normally the right thing to do. However, when someone is coming up with their wish list, that is the last thing you want. For if we had sat in that room telling them that 40 HD Independent Wireless Video Streams is just not possible they would have changed their wish list. However, because we weren’t, that is part of what the wish list brought to us.

So the teachers meet and brought us they wishlist of what they wanted.

The Wishlist:

  • 2 x ultra-short throw projectors on whiteboards in each classroom
  • 3 x wall-mounted screens (eg 50” LCDs)
  • Wireless video projection from any device to any screen (1 device to 1 screen) or from 1 device to 5 screens (one classroom) or from 1 device to 10 screens (double classroom) with a simple wall-mounted choice panel for the video projection options.
  • Hapara (https://hapara.com/) or similar to allow interactive collaboration from teacher to student

This Wishlist is exactly what I was given, it hasn’t been changed or modified for this post. In theory, it sounded simple, a bunch of screens in the room, though it was the third part that made it a little harder. They wanted to have WVP with 1:1, 1:5 and even 1:10 when you opened bi-fold doors for a double classroom, plus all manner in between.

It was at this point that we started to review what options out there we could achieve, and important want would it cost. With enough money anything is possible. WVP is something that a lot of schools had been working on for some time, we went back and reviewed posts online about WVP, we searched the MITIE Forum (https://forum.mitie.edu.au) and even posted on there asking if anyone had cracked that WVP Holy Grail.

Want we found hard was that a lot of products out there are marketing for either home networks (Apple TV/Chromecast) or big companies that want WVP for their meeting rooms. Home networks are easily, and big companies are also easy in my opinion. A user at home just wants their TV to show their phone or tablet. Maybe a few WVP in the house total. Not too hard. Businesses are also pretty easy in most cases, a series of meeting rooms that allow for WVP, but for 1 device at a time. And it is normally only with static PowerPoint slides. Plus in both scenarios, the home and or business normally have device control.

And as a lot of those reading this would know, school networks are very complicated, and when you roll something out you aren’t doing it small. For example, a WVP school rollout is normally to many locations, and your users can be in the thousands.

 

The Mesh… bring on the Faraday cage

Let’s go back to talking about technology if you remember from back to Is my wireless strong enough for WVP? I talked about how I would need two AP’s pre-room to handle the potential of 4 Vivi’s in a classroom.

A simple decision, adding another AP to the room isn’t too hard to achieve, add some more network ports, and factor that into my switch port allowance for the building and it’s done. I wasn’t worried about the cost of the AP’s, one of the benefits I loved about choosing Ubiquiti for our design was that even 2 APs cost me less than 1 of the more expensive brands. And I love saving money for the classroom.

So I now knew that I was going to put 23 AP’s in thirteen rooms. And for those wireless guys in the room how many 40 MHz double bonded channels can you get on the 5 GHz spectrum without co-channel interference? (Read more about co-channel interference here (https://www.networkcomputing.com/wireless-infrastructure/reducing-wifi-channel-interference/404994487).

The answer… 10. And with 12 rooms, 1 AP pre-room and only 10 channels I would only need to use the same channel twice, that’s fine I could put them on other sides of the building.

BUT, I wanted to put 2 APs pre-room so I actually needed to fit 24 channels where only 10 channels would fit. Another problem. I couldn’t just drop them all done to 20 MHz channels as that would still give me the same problem, not enough bandwidth to handle the WVP load. And I would have needed to go to 4 AP’s pre-room. We did also think about going up to 80 MHz channels. But again the same issue, not enough channels to go around in a close space.

In Figure 9 you can see the building layout I am dealing with, two stories one long structure with a concrete floor separating the levels.

Figure 9 – A simple 10 classroom with 2 staffroom layout.

To be honest I was unsure how best to handle this. It was our Director of ICT David Soede who went out and found a NASA white paper (https://ntrs.nasa.gov/search.jsp?R=19970036055) on signal dampening. To test this we went and brought some samples of stainless steel and copper mesh. We reviewed which ones gave us a better drop in signal strength and concluded, the gold mesh was the best. But try as we might getting Gold Mesh was just not cost effective, so we settled on the copper mesh. You can read more about how he worked it out here. https://forum.mitie.edu.au/t/faraday-cage-for-high-density-wifi/3133

 

Figure 10 – The copper mesh goes in!

While the building was all pulled apart for its refurb we placed copper mesh on the walls and roof in multiple sections of the building. The mesh was installed in big long rolls, and each sheet was connected to another via small electrical cables to allow us to earth it, and any punctures were tapped up with conductive tape.

Figure 11 – The mesh installs locations.

You can see above the yellow showing where we mounted the copper mesh, we made one mistake in this design. We didn’t think about the roof space and a metal (Colourbond) roof. Because our copper mesh didn’t go up all the way signals could bounce back and forth in the roof space, meaning we got a little signal bleed between B7, B8, B9 and B10. Other than that, we got near perfect signal block. There was one window on the bottom level that was getting signal from the top, however, it was a level so minimum that only out Wireless Expert Peter Figar from Wireless Edge Networks who found it with his wireless site survey tools, and at levels most devices wouldn’t even notice it.

Figure 12 – A  of wireless signal blockage in effect.

 Below are some bullet points from the wireless survey that was performed after the copper mesh had been installed.

B Block Ground Floor

  • Only 1x WAP from second floor detected. Strongest signal detected was -84dBm, detected from window. Most of the ground floor (represented in orange) was detected at -100dBm which most devices will not be able to detect. (see attached heat map). The rest of the floor was undetected (areas not coloured).
  • Cross floor was less effective (see ground floor – cross heat map). Heat map shows how signal drops between horizontal wall.
  • No other second floor WAPs were detected.
  • All other interference was sourced from buildings adjacent to B block.

 B Block First Floor

  • Only 1x WAP was ground floor was detected – see attached heat map. Strongest signal detected was -83dBm shown in yellow.
  • Cross floor was less effective (see first floor – cross heat map). Heat map shows how signal drops between horizontal wall.
  • No other first floor WAPs were detected.
  • All other interference was sourced from buildings adjacent to B block.

Estimate approximately 60dB loss from concrete+copper shielding between floor. Could be more since survey tablet cannot read below -100dBm. For example, maximum AP signal strength detected on ground floor was -37dBm; same AP was not detected on First floor. Lowest signal readable by survey tablet is -100dBm. 100 – 37= 63dB. Estimate about 25-30dB loss between floors at I Block.

 

Summary benefits of HP WVP

Below is the list of benefits you get with a well designed and implemented HD WVP system.

For teachers;

  • Teachers don’t have to plan multi-screen WVP activities based on content type eg static vs HD Video.
  • Can be flexible with pedagogy & activities. Eg mid-lesson they can swap to simultaneous multiple HD WVP without worrying about will it work.
  • Untethered, can see student groups work simultaneously no matter content type.

Students;

  • Can use local content from their phone, tablet or notebook, not just YouTube or ClickView sourced content that relies on the handoff feature in Vivi or Apple TV.
  • No hardware dependency based on ports or dongles etc. As we are BYOD year 10 to 12 this was fantastic.

ICT Department;

  • More flexibility when choosing devices for our staff or students (Years K-9, school supplied device.)
  • Future proofs for growth in wireless usage (due to increased video or internet usage etc)

 

Teacher and Engagement

When rolling out something new it always comes with potential issues, a big one I have found in schools is that users are happy with what they have. Unless someone can show them the educational benefits of changing the staff just don’t want to. And why should they? They have a limited time to teach the students and they don’t want to waste time learning a new way to teach if they see no benefit. And to some staff introducing Vivi is a nice addon, though the HDMI cable worked just fine to them. And those who aren’t in the fancy new B Block would have seen it as more of a gimmick than a benefit when their room just had 1 screen and a standard classroom layout.

For CCGS though, the nature of our gave us two benefits that helped achieve a successful rollout.

  • When the Vivi’s were being installed, we changed the setting of all our Projectors. Whereas in the past the HDMI cables were the default option when a projector was turned on. Now we had Vivi’s as the default. This meant that when a teacher came into a room they had to first turn the projector on, and then select Vivi. This meant it was an extra step for staff.
  • We also rolled out our new staff notebook devices, the Dell 7285 which is a convertible tablet-style notebook. And due to its slim size and only have USB-C ports it meant staff also had to carry a USB-C to HDMI adapter with them.

Now yes part of this was ‘forcing’ them to adapt, however, any staff that didn’t want to could get an HDMI adapter and knew to charge their projector each time. For example, our Junior Staff wanted the interactive functionality that Vivi at the time didn’t offer*. So they would make the changes.

* When we first rolled out Vivi interactivity was not an option they offered, it has since been released.

Because of this, we saw the great update, you can see in the below pictures pulled out of our Vivi console the high level of use we are getting. We even have some staff who connect Monday morning to the Vivi and they don’t disconnect until they shut down their notebook on Friday afternoon.

Figure 13. Teacher and Student Engagement 1

Figure 14. Top Staff and Student usage in Hours over a month.

Crucial Aspects of doing WVP well

I am starting to wrap up here, here is a list of the area’s we found were crucial to getting WVP done well.

  • Your Wireless.
  • 802.11ac 5GHz with 40 or 80 MHz channels. Good channel mapping.
  • Limit the number of devices per Wireless AP
  • Have load balancing on bandwidth or utilisation.
    (not just load balancing based on numbers)
  • Wired network to VVP devices.
  • Gigabit to each AP.

 

Vivi Agent… quick run-through

Quickly, I felt it worth showing the ease of use for the Vivi Agent, in Figure 15 you see the steps taken to select your building, room and then present your screen to that room. They are 2 additional steps when using AirPlay.

Before I move on from Vivi, I just wanted to show the easy steps that any staff member can take to connect to the screens, from opening up the App they only have 5 steps to go from selecting the building they into presenting. If they are on a Mac it has 1 extra step of selecting the room from the AirPlay drop down.

Figure 15. Vivi Present my screen steps, bottom 2 are just when on a Mac.

And in Figure 16 you can see the easy 3 steps in handing off a YouTube video to be played. The benefit of this is that it doesn’t require any high bandwidth on your wireless as the Vivi box itself is playing the video, not mirroring your screen playing the video across your wireless.

Figure 16. Vivi Handoff of Youtube.

Finally, in Figure 17 and 18 you can see the room layout and presenting your screen. In Figure 17 we see my MacBook Pro seating in there, I have the Vivi agent up but I haven’t presented yet.

Then in Figure 18, you can see the screens are all mirroring mine after following the above steps. And of course, I can choose how many screens I wanted to present to, in this one I choose them all.

NOTE: Sorry for the crappy picture quality, I had to use panorama to get everything in this picture.

Figure 17, Vivi’s ready to connect.

Figure 18. My MacBook pro screen now presenting to all the Vivi’s.

And as you can see in Figure 19 below our rooms turned out well. Each room got its two projectors and an additional one or two 65″ Dell HD screens depending on the layout. We did want to get more Dell HD screens around the rooms, however, we needed to make sure that we installed them with enough room around them for students to work effectively. Around the Dell Screens are glass whiteboard panels the kids can write on, these are the Digital Collerabation Zones of DCZ I mentioned earlier.

Figure 19. The completed double classroom.

Conclusion

In wrapping up I come back to the original statement that was made. 40 HD Independent Wireless Streams in one building… a world first.

Due to the way Vivi works, we have a teacher in each of the rooms pushing their screen to all the available screens in their room as 12x 1:3/1:4. Or we have each screen in the room pushing their own content 40:40. David Soede and I have done a few talks at some Australian conferences around this topic, and so far, no one has been able to challenge us around our statement. If you can prove me wrong I would be happy to know.

In Video 1, you can see how these rooms are used on a daily bases by our staff and students.

Video 1 – A walk through a double classroom.

And I’ll end the post here, with the teachers from CCGS talking about their new rooms and how they feel it has changed the way they do teaching and learning.

Video 2 – Next Generation Learning Spaces – Teacher Insights

Leave a Reply

Your email address will not be published. Required fields are marked *