In the past few years, the needs of a wireless network have evolved greatly. Early on, access to a wireless network at a school was a convenience that allowed a handful of mobile devices access to the internet. With the lack of mobile devices and the financial investment made on wired computer labs, the wireless network was far from mission critical. Today, other than sports stadiums and some hospitals, schools have the highest client density and some of the most demanding wireless needs of any user group. Schools now have a wireless device, typically a laptop or tablet, for every student and teacher in the district. The wireless network must also sometimes support a wide range of cell phones and other random devices such as printers, A/V equipment, and other devices of varying wireless capabilities. With the influx of wireless devices, the thought process in which a wireless network was designed also had to shift from providing coverage designs to engineering capacity designs.
Coverage – to steal words from Keith Parsons, founder of Wireless LAN Professionals, “coverage is easy.” A “coverage design” is essentially providing enough of an RF signal for a wireless client to associate to an access point and transmit and receive data. This model uses the minimum number of access points possible to cover a space. It does not take client capabilities or application throughput into consideration. If your device can see the wireless network and connect, the coverage design is working perfectly.
Capacity – this is a much more in depth and detailed design that is required for medium to high client density environments. In environments such as these, everything must be done to allow the wireless clients to transmit and receive as quickly as possible so that the channel air time does not get overutilized. Each channel has a finite amount of airtime in which a client or AP can transmit/receive. The amount of airtime available is dictated by numerous variables, such as the number of SSIDs being broadcast (more SSIDs = more management overhead), PHY types supported (802.11a,b,g,n,ac), number of spatial streams the wireless clients and AP support and numerous other variables. The amount and the complexity of the variables involved is one reason to engage a wireless engineer when upgrading a current wireless infrastructure or the needs of the existing infrastructure have changed. It may seem like a logical assumption that you can place new access points where the old ones are, but each access point radiates energy in different patterns. The back corner where the teacher’s desk is could easily have less RF energy (or signal strength) just by changing access point models. Less RF energy equals slower data rates or, in severe cases, association issues due to lack of signal strength.
Environment – the physical environment plays a huge role in RF design. Different wall materials absorb (attenuates) different amounts of RF energy. Common wall materials in schools include drywall, cinder block, concrete filled cinder block, brick, security glass, and sometimes retractable metal walls. Each one of these materials attenuate RF energy differently. No wall is made the same – a drywall wall that attenuates 5dBm of signal at one school may only attenuate 3dBm at another school. It is important to have a wireless engineer measure wall attenuation to determine the correct power settings for your access points.
Something else to keep in mind about the environment is the current RF noise. If a school is in a residential neighborhood, it could potentially have numerous wireless networks surrounding it, using overlapping channels or large channel widths. It’s important that an RF survey is completed prior to access points being deployed so the access points can be put on the correct channels to avoid any adjacent channel interference and to minimize co-channel interference, that could potentially have a negative impact on the wireless throughput of the school. There are also other sources that can negatively impact the wireless environment – microwaves, cordless phones, wireless printers, wireless video cameras, etc. Bluetooth devices in large quantity can impact the 2.4ghz band. You may have been told that Bluetooth interferes with your wireless network, but this is not true. First, Bluetooth uses extremely low power. Second, Bluetooth uses a very small section of the 2.4ghz band. In a band where Wi-Fi is divided into 11 channels, Bluetooth breaks the same frequency range into 79 channels. Third, Bluetooth changes channel (hops) 1600 times per second. In cases where there is a lot of Bluetooth, the worst it will do is raise the noise floor. This can be detrimental to the wireless network, but it only occurs on the 2.4ghz band, and I am hoping by now you do not use the 2.4ghz band for your school owned devices.
Devices – the devices you allow on your wireless network will have a huge impact of the success of the network. Wireless devices that are on the market currently have between 1 and 3 transmitters and receivers and use between 1 and 3 spatial streams; this is where 1x1:1, 2x2:2 and 3x3:3 come from – TxR:SS. Having more transmitters, receivers, and spatial streams allow the device to have a greater receive sensitivity, which allows it to decode wireless frames better (think of them as ears, you can decipher speech better with 2 ears than 1 ear). The number of spatial streams allow the devices to receive and transfer data more quickly. The data rate at which the device transfers data is directly related to the number of spatial streams the wireless device has. Once again, the more quickly the device can transmit data, the more quickly it can get off the wireless medium and let another device use it.
The amount of devices being used in a given area should also be taken into consideration. Since wireless is a half-duplex medium (only one device can send or receive at a time – no longer the case with 802.11ax (WiFi 6)), all other devices on a given channel must wait for the device that is transmitting to complete. The more devices on a channel, the more data is transmitted and therefore more devices waiting. Think of this like a group of people speaking to one another. Everyone listens to the person talking at the time and when there is a slight silent period, another person may begin speaking. If 2 people begin speaking at the same time, no one can understand what is being said and both parties will stop and one will start again – in WiFi this is called a collision and results in a retransmission of the frames. How do you know if you have too many devices in one room for the number of access points? Like most things in WiFi, it depends! What applications are being used? What throughput is needed for the applications to function? What is the SNR the client is seeing in its area? What band are the clients using? Is there adjacent channel interference or co-channel interference? A wireless engineer can help determine the answer to these questions and help design for the specific devices and usage.
Applications – as mentioned above, planning for the applications used across the wireless network is crucial to ensuring that the network can support the number of wireless clients accessing it at a given time. In an environment where there is a 3x3:3 802.11ac access point and 30 2x2:2 802.11ac clients, if the required throughput is 1Mbps second, the area will only need 1 access point. If the required throughput is 5Mbps and the clients and access point does not change, you now need 3 access points to provide that type of throughput to all 30 wireless clients. With most things in schools being web based, needing 5Mbps of throughput per client is quite excessive, but it is something that needs to be evaluated and designed for. Typically, HD video streaming can range from 2 – 5 Mbps, while web browsing and YouTube are around 500kbps each.
No Design – this is by far the most common mistake there is. In schools, the default seems to be an access point for every classroom. Do you need an AP in every classroom? It depends. What are the wall materials? What are your devices? What are your throughput needs? Was wall attenuation measured? Was a proper predictive wireless design completed using the answers to these questions? If so, and it was determined that an AP was needed in each classroom, the answer is yes, you do need an AP in each classroom. If none of those questions were answered or even discussed, the answer could possibly still be yes, but we just don’t know. Wireless is a very resilient technology and even very poor wireless networks still work, they just don’t work well. Because of this resiliency, poor performance is masked for years until client density increases to a point where the issues become very evident. A wireless engineer can view the RF environment and see issues you cannot. They have the ability to capture wireless frames and see issues you don’t know you have and may be able to fix issues before they appear.
Everything is wireless – the rule here is very simple – if it moves, it’s wireless. If it doesn’t move, run a wire. Media devices, such as Apple TV and projectors, rarely or never move. These devices can create a heavy load on a wireless network due to the nature of their use – streaming video. Due to this and the fact they never move, it makes more sense to run a network cable to these devices and save the valuable channel airtime for the devices that are critical to the education of the students. Another device that falls under this category are printers. Not only can they operate on a wireless network, but they can also have their own ad-hoc network that it beacons. Not only does a printer not move but if the ad-hoc network is left enabled, it will function according to the 802.11 standard and beacon every 100ms on the channel assigned. This once again takes away the most precious thing in wireless – airtime.
Excessive SSIDs – the effect of excessive SSIDs can be very negative impact on your wireless network. Each SSID comes with a significant amount of management overhead that must be accounted for. Each SSID, even hidden SSIDs, send out a beacon every 100ms. The beacon frame is a very large frame that does a number of jobs, which include advertising the network and AP capabilities and indicating to clients in power save mode when traffic is queued for them. More SSIDs = more beacons. More beacons = less airtime available for clients to transmit. Not only are beacon frames large, they are also sent at the slowest data rates, 1Mbps in the 2.4ghz band and 6Mbps in the 5ghz band, for backwards compatibility. It’s commonly stated that best practice is to have no more than 3 SSIDs. There are a lot of factors that go into figuring this out, but 3 is a good place to start. Typically, if you have more than 3, you may not truly understand how SSIDs function.
A common misconception with SSIDs is that they function like a VLAN. This is very far from the truth. The SSID can be assigned to a specific VLAN to segregate traffic once on the wire, but while the traffic is modulated over the air, it is not separated and can be heard regardless of SSID or VLAN assigned to it. Due to this misconception, I have seen several instances where SSIDs are created for specific devices or specific rooms, BUT they all place the traffic on the same VLAN and the settings for each SSID are the exact same. This does nothing but create more management overhead and degrade the performance of your wireless network.
If your devices do need to be segregated on different VLANs, the best way to accomplish this is with the use of a RADIUS server. You will then be able to create a single SSID, and the device can be dynamically moved to the correct VLAN depending on the RADIUS attributes that are returned.
Wireless networks are very complex systems that require a lot of expertise to plan, design, and implement. There is a stark difference between selling a solution and engineering a solution. Anyone can come in and use a few buzzwords and tell you an AP in every classroom is the way to go and in some cases, they may be correct, but can they explain WHY you need an AP in every classroom? Unfortunately, with E-Rate and the bidding systems schools are constrained to, this is very common and doesn’t allow for design work to be performed. The fact is, you will engage a wireless engineer at some point. The question is, will it be to design a quality wireless network or when you have issues that arise from a wireless network that was not designed? The cost and time spent will be far greater to troubleshoot the network than to design it properly from the beginning.
Schools have quickly become a high-density client device environment and the implementation of online testing and the investment in wireless devices has put a lot of pressure on school IT departments to get the wireless network working correctly. Schools systems that have engaged a wireless engineer to plan their network are having great success with online testing and 1:1 device deployment.
If you have questions and/or concerns about the wireless network design in your school or business, engage our team. We’d be happy to do an assessment or have a discussion.