Why must we design Wi-Fi? Why can’t we just install access points and call it a day? That’s something I go into with this webinar. But the short story is you don’t want to guess. We’re delivering a service that is used by thousands of people and whom rely on this service.
And some of those people have over 30k clients per day. What if 1/3 of those people had poor Wi-Fi experiences and put in a ticket for each incident?
When it comes to Wi-Fi, we should follow a process or a lifecycle in order to continue delivering a good user experience. I go into each of these on the webinar.
I provide examples from my own environments.
Share my solutions to temporary Wi-Fi deployments for events.
Historically, AP-on-a-Stick (APoS) surveys were the only way to find out where to place the access points. We didn’t have any of the fancy software we have today to predict how the Wi-Fi signal will behave in our space. However, we still perform APoS today. In this episode, we will explain why we still use them and how they can be beneficial to us Wi-Fi Engineers.
APoS utilizes an AP that will be deployed on-site. The AP would be installed on a “stick” and is used to measure how the Wi-Fi signal behaves in the space that is being surveyed. The advantages are it takes into consideration the behavior of an RF wave and the attenuation, reflection, diffraction, scattering, etc. The disadvantage is it takes time and materials to achieve.
Historically, that was the way we used to define where to place the APs. We would start in a corner of a building and place the first AP, measure the signal to define the edge of the Wi-Fi cell and go on to the next measurement until you covered the whole site and defined the different AP locations.
The introduction of tools allowed us to model and predict how the Wi-Fi signal will propagate in a given space. This allows us to plan and perform our Wi-Fi design “on paper” so to speak. In order for this type of predictive design to be accurate, we still need to understand the physical environment. However, we can predict the signal prediction in the software and save some time.
Why do we still use APoS then? In some environment and cases, the predictive design will not be accurate enough for us to be confident in the design. In these cases, we now use the APoS to validate some of the AP locations defined during the predictive design phase. How much validation do we need to do? It depends 🙂
What material will you require to perform an APoS? The APs and antennas you will be using for your deployment. If you are not too sure which AP or Antenna will be used, test multiple models. A stick (tripod, poles…). If you need to go very high, you might need a scissor lift or other heavier tools. A bracket to install the AP and antenna on (Wi-Fi Stand). If you are planning on using specific angles, it could be useful to have a bracket that displays the angle. A battery to power on the AP. Venvolt from Ventev, Accelerator from AccelTex, A long ethernet cable to connect the battery to the AP (I like the flat ones), A water bottle, Useful add-ons: flashlight, Duck tape, zip ties, marker, security boots, yellow vest, ear plugs, safety glasses.
How do you prepare an APoS? You will need to configure the access point. Configure the access point in a standalone mode. Each vendor has their own way of doing it. Cisco using Mobility Express, Aruba using Instant AP, Meraki using the Survey Mode.
I like to configure 1 SSID per band so that I can analyze the results in an easier matter afterward. It also helps me to make sure I am walking far enough to reach the edge of both frequency cell edge. I use the transmit powers used for the design. You can use the channels you want. In order to plan for the worst case scenario, you could choose the highest channel you can use in your space. (Channel 11 on 2.4GHz and Channel 165 on 5GHz). You should also have your predictive model ready before going onsite for the APoS measurements (as much as you can).
How do you perform an APoS? How do you choose which AP location to test? Place the APoS exactly how you would install the AP in production. Turn on the AP and wait for the SSID to be broadcasted. Use your favorite survey software to measure the signal in passive mode. Keep an eye on the signal strength you receive and make sure you walk past your threshold. Once you are done, freeze your AP (in Ekahau) before you move the AP to the next measurement. Try to document your work as you go by taking pictures of the AP at its measurement location and put in notes in your site survey software.
Analyze the results and compare how the signal propagates in the real world and what you had in the predictive design. Adjust your model so it matches the real world. It will be an iterative process and both can be done on-site while you do your APoS.
Things you can also do while you are on-site: Validate the AP locations will work for all the location you have defined in your predictive model. Mark the AP location if you want to. Test the different options such as height and antennas. Take a look at how the users use Wi-Fi.
Rowell and François were invited to Cisco Headquarters, along with many others, as part of the #JustTheTech event which was planned in coordination with the Wi-Fi 6 announcement that happened on April 29, 2019.
It was a jam packed day of discussions with key people from Cisco.
We kicked off the event with a tour of the Customer Experience Center. Taking a look at the possibilities using Cisco solutions. It’s an impressive facility on the first floor of one of many Cisco buildings. There’s an entry way with a large Cisco logo on the wall. On the other side of the wall is the beginning of the Customer Experience Center.
Big Cisco logo
Hop aboard the Cisco bus
Virtual Concierge
Industrial switches
The Customer Experience Wall
Cisco’s first product, the AGS.
BGP on a napkin
During our tour we ran into Todd Nightingale who is an SVP, General Manager of Meraki. It was a delightful surprise but he was able to talk to us for a bit and even tour with us.
So the main purpose of #JustTheTech was to take part in the coordinated announcement of Wi-Fi 6 products from both Cisco and Meraki. But instead of marketing material, we were able to speak directly to engineers.
This means taking a look at the new Wi-Fi 6 access points up close
Catalyst 9115, 9117, 9120
Meraki MX55, MX45
The actual event started off with Sacha Gupta, Senior Vice President, Product Management – Cisco Enterprise Networking. He talked about Reinventing Access, Unplugged and Uninterrupted.
Reinventing Access, Unplugged and Uninterrupted by Sacha Gupta
Next, Jérôme Henry, Principle Engineer, Office of the CTAO, talked about OpenRoaming which is an interesting concept in today’s world where 5G and Wi-Fi 6 will converge. Cisco has developed partnerships with the biggest device manufacturers showing they really want to make these products work on our networks.
Understanding OpenRoaming and the Device Ecosystem Project by Jérôme Henry
Next, Fred Niehaus and Jeevan Patil begin their combined talk about Wi-Fi 6. This is where we really get interested as everyone was expecting new Wi-Fi 6 hardware.
What Wi-Fi 6 will bring to the market by Fred Niehaus & Jeevan Patil
Of course, Fred goes into the details of the new Catalyst access points. These are the equivalent of the AireOS 1800 and 2800 access points.
Introducing Cisco’s Catalyst Wi-Fi 6 Access Points by Fred Niehaus
Jeevan introduces Meraki’s new Wi-Fi 6 access points. They look radically different from the Cisco Catalyst access points. Something out of a Marvel movie.
Next-Generation Meraki Access Portfolio by Jeevan Patil
And as a bonus, we get to hear about the new Catalyst 9600 chassis switch. The thing is a beast.
Bedrock of the Cloud-scale Campus by Shawn Wargo
Speaking of the Catalyst 9600 Series, there was a Roundtable Discussion.
We also took part in a Wi-Fi 6 Roundtable Discussion with Fred Niehaus and Cristian Raducanu.
One of the highlights of the day was taking a shuttle over to another Cisco building where Wi-Fi 6 access points and devices are tested. This facility is essential as all of the tests for Wi-Fi 6 starts here. Inside the building is a caged off room – it literally has a chain-link fence which requires special badge privileges.
There are tables of laptops with Wi-Fi 6 adapters installed and also many Wi-Fi 6 mobile devices. Currently, being all Samsung as they were the first ones out with Wi-Fi 6. On the ceiling are the Catalyst 9115, 9117, or 9120 access points.
Various tools were used to test Wi-Fi 6 connectivity and roaming. We had a question of 1024 QAM come up and the test engineers stated it was possible to get 1024 QAM at some distance. No access point above your head required.
Closing Thoughts
We were fortunate to be invited to Cisco’s headquarters and take part in this awesome event. It does help that we are both Cisco Champions. With a small group of individuals, it was an intimate setting. The access we had to engineers within Cisco was great and led to great discussions about Wi-Fi 6 including an up-close look at the products.
With Wi-Fi 6 and 5G now taking the stage, one has to think – what about the coexistence of the two technologies? Rowell and François met together at Cisco HQ to speak with Jérôme Henry about OpenRoaming and Cisco’s vision of seamless roaming.
François Vergès (left), Jerome Henry (center), Rowell Dionicio (right)
Rowell and François participated at an event called #JustTheTech at Cisco HQ in Milpitas, CA. At this event we were table to speak to technical experts just about the technology and how it works in regards to Wi-Fi 6, switching, testing, and more.
In this episode, François talks with Jérôme Henry, following his presentation on OpenRoaming. In the room is Rowell Dionicio, Sam Clements, and Dave Benham – all who participated at #JustTheTech.
OpenRoaming
What is OpenRoaming? According to Cisco:
Cisco, along with other vendors and enterprises, is working to provide a better bridge between mobile devices and Wi-Fi networks. With frictionless and secure guest onboarding, users can roam across Wi-Fi 6 and 5G networks, automatically maintaining connectivity with security and achieving the ultimate experience.
Evolution of the modulation techniques we are using today with 802.11ax (256-QAM).
With 1024-QAM, we are now able to encode 10 bits per cycle on each subcarrier. The way we are able to do that is by increasing the number of different levels of amplitudes used to encode the data.
If you want to learn more about the different types of modulations used by Wi-Fi and how they work, there is a great video where Keith Parsons explains it on youtube: https://www.youtube.com/watch?v=W5DMfEuY2Vg&t=8s
Due to the addition of a new modulation technique (QAM-1024), 2 new MCS indexes are now available with 802.11ax:
Index 10: when the 1024-QAM modulation is used with a coding of 3/4
Index 11: when the 1024-QAM modulation is used with a coding of ⅚
FEC (Forward Error Correction). Send more than the data bits. If you lose some of the sequence, the remaining bits will help you to understand what you were supposed to be sent.
What is the challenge with more complex modulation techniques?
How well will 1024-QAM in real life?
Will we be able to take advantage to it?
Interesting talk on twitter around the subject (Troy Martin, Andrew, Hendrik Lüth & Jim Vajda). We tend to think that a smaller communication bandwidth will give us a better SNR. But it is not necessarily the case here since the receiver in 802.11ax will still be listening to the whole 20MHz wide channel even if its RU is smaller. Link here: https://twitter.com/VergesFrancois/status/1113779145731977216
With 802.11ax, we are getting a whole new set of data rates (or MCSs). If we want to understand why, we need to understand how these data rates are calculated.
The amount of data we can transfer through a Wi-Fi link will depend on:
The channel width (or the number of subcarriers)
The modulation and coding use
The amount of spatial streams used
The guard interval used
The duration of the symbol
And we can actually take all these different variables and calculate the different data rates using the following formula:
802.11n/ac Data Rate Formula
You can take a look at the blog post and see which values can each of these variables can have for 802.11n and 802.11ac:
HT & VHT Parameters
Now, the reason why we have a new set of data rates for 802.11ax is because some of the key variables are changing:
A new symbol duration is used: 12.8µs
Different Guard Intervals are used: 0.8µs, 1.6µs and 3.2µs
The size and number of data subcarriers is not the same (especially with the different RU sizes introduced by ODFMA.
Also, with the introduction of OFDMA and the use of Resource Units, we might be using smaller numbers of subcarriers which impact the data rates.
This is why the draft identify the OFDMA and non-OFDMA MCS differently.
The draft even use specific variables related to each resource unit. And we can therefore define this new formula:
802.11ax ODFMA Data rate Formula
And here are the different values each of these variables can have for 802.11ax communications. The first table details the parameters used when OFDMA is not used. The second table details the parameters when OFDMA and resource units are used.
802.11ax OFDM Parameters
802.11ax OFDMA Parameters
Sections Talking about MCS in The Standards & Draft
28.3.7 HE Modulation and coding schemes (HE-MCSs) (p.442)
